* [PATCH 00/18] xfsprogs: refactor buffer function names
@ 2020-02-20 1:42 Darrick J. Wong
2020-02-20 1:42 ` [PATCH 01/18] libxfs: clean up readbuf flags Darrick J. Wong
` (17 more replies)
0 siblings, 18 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:42 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
Hi all,
This series cleans up several messes in the libxfs buffer handling code.
First, we get rid of the overloaded (and in some places hidden usages)
of LIBXFS_EXIT_ON_FAIL flag that is sprinkled throughout the buffer
callers. Next, we rename the buffer get/read/put/write functions to
match their kernel counterparts, which enables us to remove a bunch of
ugly #defines. Then, we replace the open-coded uncached buffer logic in
the callers with the same uncached buffer API as the kernel has.
Finally, we move as many callers as we feasibly can to use the
xfs_buf_(get|read) interfaces so that we don't have multiple entry
points to the same functionality.
If you're going to start using this mess, you probably ought to just
pull from my git trees, which are linked below.
This is an extraordinary way to destroy everything. Enjoy!
Comments and questions are, as always, welcome.
--D
xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=libxfs-refactor-buffer-funcs
^ permalink raw reply [flat|nested] 44+ messages in thread
* [PATCH 01/18] libxfs: clean up readbuf flags
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
@ 2020-02-20 1:42 ` Darrick J. Wong
2020-02-21 14:42 ` Christoph Hellwig
2020-02-20 1:42 ` [PATCH 02/18] libxfs: clean up writebuf flags Darrick J. Wong
` (16 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:42 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Create a separate namespace for libxfs_readbuf() flags so that it's a
little more obvious when we're trying to use the "read or die" logic.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/init.c | 11 ++++++++---
libxfs/libxfs_io.h | 3 +++
libxfs/rdwr.c | 4 ++--
mkfs/xfs_mkfs.c | 4 ++--
4 files changed, 15 insertions(+), 7 deletions(-)
diff --git a/libxfs/init.c b/libxfs/init.c
index d1d3f4df..428497f0 100644
--- a/libxfs/init.c
+++ b/libxfs/init.c
@@ -640,6 +640,7 @@ libxfs_mount(
xfs_buf_t *bp;
xfs_sb_t *sbp;
int error;
+ int readflags = 0;
libxfs_buftarg_init(mp, dev, logdev, rtdev);
@@ -716,9 +717,13 @@ libxfs_mount(
if (dev == 0) /* maxtrres, we have no device so leave now */
return mp;
+ /* device size checks must pass unless we're a debugger. */
+ if (!(flags & LIBXFS_MOUNT_DEBUGGER))
+ readflags |= LIBXFS_READBUF_FAIL_EXIT;
+
bp = libxfs_readbuf(mp->m_dev,
d - XFS_FSS_TO_BB(mp, 1), XFS_FSS_TO_BB(mp, 1),
- !(flags & LIBXFS_MOUNT_DEBUGGER), NULL);
+ readflags, NULL);
if (!bp) {
fprintf(stderr, _("%s: data size check failed\n"), progname);
if (!(flags & LIBXFS_MOUNT_DEBUGGER))
@@ -733,7 +738,7 @@ libxfs_mount(
(!(bp = libxfs_readbuf(mp->m_logdev_targp,
d - XFS_FSB_TO_BB(mp, 1),
XFS_FSB_TO_BB(mp, 1),
- !(flags & LIBXFS_MOUNT_DEBUGGER), NULL))) ) {
+ readflags, NULL))) ) {
fprintf(stderr, _("%s: log size checks failed\n"),
progname);
if (!(flags & LIBXFS_MOUNT_DEBUGGER))
@@ -760,7 +765,7 @@ libxfs_mount(
if (sbp->sb_agcount > 1000000) {
bp = libxfs_readbuf(mp->m_dev,
XFS_AG_DADDR(mp, sbp->sb_agcount - 1, 0), 1,
- !(flags & LIBXFS_MOUNT_DEBUGGER), NULL);
+ readflags, NULL);
if (bp->b_error) {
fprintf(stderr, _("%s: read of AG %u failed\n"),
progname, sbp->sb_agcount);
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index fc0fd060..b294e659 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -124,6 +124,9 @@ extern struct cache_operations libxfs_bcache_operations;
#define LIBXFS_GETBUF_TRYLOCK (1 << 0)
+/* Exit on buffer read error */
+#define LIBXFS_READBUF_FAIL_EXIT (1 << 0)
+
#ifdef XFS_BUF_TRACING
#define libxfs_readbuf(dev, daddr, len, flags, ops) \
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 92e497f9..32619a8d 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -911,13 +911,13 @@ __read_buf(int fd, void *buf, int len, off64_t offset, int flags)
int error = errno;
fprintf(stderr, _("%s: read failed: %s\n"),
progname, strerror(error));
- if (flags & LIBXFS_EXIT_ON_FAILURE)
+ if (flags & LIBXFS_READBUF_FAIL_EXIT)
exit(1);
return -error;
} else if (sts != len) {
fprintf(stderr, _("%s: error - read only %d of %d bytes\n"),
progname, sts, len);
- if (flags & LIBXFS_EXIT_ON_FAILURE)
+ if (flags & LIBXFS_READBUF_FAIL_EXIT)
exit(1);
return -EIO;
}
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 6b182264..a57046f1 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -3577,7 +3577,7 @@ rewrite_secondary_superblocks(
XFS_AGB_TO_DADDR(mp, mp->m_sb.sb_agcount - 1,
XFS_SB_DADDR),
XFS_FSS_TO_BB(mp, 1),
- LIBXFS_EXIT_ON_FAILURE, &xfs_sb_buf_ops);
+ LIBXFS_READBUF_FAIL_EXIT, &xfs_sb_buf_ops);
XFS_BUF_TO_SBP(buf)->sb_rootino = cpu_to_be64(mp->m_sb.sb_rootino);
libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
@@ -3589,7 +3589,7 @@ rewrite_secondary_superblocks(
XFS_AGB_TO_DADDR(mp, (mp->m_sb.sb_agcount - 1) / 2,
XFS_SB_DADDR),
XFS_FSS_TO_BB(mp, 1),
- LIBXFS_EXIT_ON_FAILURE, &xfs_sb_buf_ops);
+ LIBXFS_READBUF_FAIL_EXIT, &xfs_sb_buf_ops);
XFS_BUF_TO_SBP(buf)->sb_rootino = cpu_to_be64(mp->m_sb.sb_rootino);
libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
}
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 02/18] libxfs: clean up writebuf flags
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
2020-02-20 1:42 ` [PATCH 01/18] libxfs: clean up readbuf flags Darrick J. Wong
@ 2020-02-20 1:42 ` Darrick J. Wong
2020-02-20 1:42 ` [PATCH 03/18] libxfs: remove LIBXFS_EXIT_ON_FAILURE Darrick J. Wong
` (15 subsequent siblings)
17 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:42 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Create an explicit namespace for writebuf flags so that we can keep the
flags namespaces separate.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 3 +++
libxfs/rdwr.c | 11 +++++++++--
mkfs/proto.c | 2 +-
mkfs/xfs_mkfs.c | 14 +++++++-------
4 files changed, 20 insertions(+), 10 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index b294e659..bb6b689e 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -127,6 +127,9 @@ extern struct cache_operations libxfs_bcache_operations;
/* Exit on buffer read error */
#define LIBXFS_READBUF_FAIL_EXIT (1 << 0)
+/* Exit on buffer write error */
+#define LIBXFS_WRITEBUF_FAIL_EXIT (1 << 0)
+
#ifdef XFS_BUF_TRACING
#define libxfs_readbuf(dev, daddr, len, flags, ops) \
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 32619a8d..f56303e2 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -1177,21 +1177,28 @@ libxfs_writebuf_int(xfs_buf_t *bp, int flags)
}
int
-libxfs_writebuf(xfs_buf_t *bp, int flags)
+libxfs_writebuf(
+ struct xfs_buf *bp,
+ int flags)
{
+ int bflags = LIBXFS_B_DIRTY;
+
#ifdef IO_DEBUG
printf("%lx: %s: dirty blkno=%llu(%llu)\n",
pthread_self(), __FUNCTION__,
(long long)LIBXFS_BBTOOFF64(bp->b_bn),
(long long)bp->b_bn);
#endif
+ if (flags & LIBXFS_WRITEBUF_FAIL_EXIT)
+ bflags |= LIBXFS_B_EXIT;
+
/*
* Clear any error hanging over from reading the buffer. This prevents
* subsequent reads after this write from seeing stale errors.
*/
bp->b_error = 0;
bp->b_flags &= ~LIBXFS_B_STALE;
- bp->b_flags |= (LIBXFS_B_DIRTY | flags);
+ bp->b_flags |= bflags;
libxfs_putbuf(bp);
return 0;
}
diff --git a/mkfs/proto.c b/mkfs/proto.c
index 2ece593e..0025fa08 100644
--- a/mkfs/proto.c
+++ b/mkfs/proto.c
@@ -262,7 +262,7 @@ newfile(
if (logit)
libxfs_trans_log_buf(tp, bp, 0, bp->b_bcount - 1);
else
- libxfs_writebuf(bp, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(bp, LIBXFS_WRITEBUF_FAIL_EXIT);
}
ip->i_d.di_size = len;
return flags;
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index a57046f1..db640a11 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -3447,7 +3447,7 @@ prepare_devices(
buf = libxfs_getbuf(mp->m_ddev_targp, (xi->dsize - whack_blks),
whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
libxfs_purgebuf(buf);
/*
@@ -3458,7 +3458,7 @@ prepare_devices(
*/
buf = libxfs_getbuf(mp->m_ddev_targp, 0, whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
libxfs_purgebuf(buf);
/* OK, now write the superblock... */
@@ -3466,7 +3466,7 @@ prepare_devices(
buf->b_ops = &xfs_sb_buf_ops;
memset(buf->b_addr, 0, cfg->sectorsize);
libxfs_sb_to_disk(buf->b_addr, sbp);
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
libxfs_purgebuf(buf);
/* ...and zero the log.... */
@@ -3486,7 +3486,7 @@ prepare_devices(
XFS_FSB_TO_BB(mp, cfg->rtblocks - 1LL),
BTOBB(cfg->blocksize));
memset(buf->b_addr, 0, cfg->blocksize);
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
libxfs_purgebuf(buf);
}
@@ -3579,7 +3579,7 @@ rewrite_secondary_superblocks(
XFS_FSS_TO_BB(mp, 1),
LIBXFS_READBUF_FAIL_EXIT, &xfs_sb_buf_ops);
XFS_BUF_TO_SBP(buf)->sb_rootino = cpu_to_be64(mp->m_sb.sb_rootino);
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
/* and one in the middle for luck if there's enough AGs for that */
if (mp->m_sb.sb_agcount <= 2)
@@ -3591,7 +3591,7 @@ rewrite_secondary_superblocks(
XFS_FSS_TO_BB(mp, 1),
LIBXFS_READBUF_FAIL_EXIT, &xfs_sb_buf_ops);
XFS_BUF_TO_SBP(buf)->sb_rootino = cpu_to_be64(mp->m_sb.sb_rootino);
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
}
static void
@@ -3939,7 +3939,7 @@ main(
if (!buf || buf->b_error)
exit(1);
(XFS_BUF_TO_SBP(buf))->sb_inprogress = 0;
- libxfs_writebuf(buf, LIBXFS_EXIT_ON_FAILURE);
+ libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
/* Make sure our new fs made it to stable storage. */
libxfs_flush_devices(mp, &d, &l, &r);
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 03/18] libxfs: remove LIBXFS_EXIT_ON_FAILURE
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
2020-02-20 1:42 ` [PATCH 01/18] libxfs: clean up readbuf flags Darrick J. Wong
2020-02-20 1:42 ` [PATCH 02/18] libxfs: clean up writebuf flags Darrick J. Wong
@ 2020-02-20 1:42 ` Darrick J. Wong
2020-02-20 1:43 ` [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse Darrick J. Wong
` (14 subsequent siblings)
17 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:42 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Now that all the implicit users of this flag are gone, remove its
definition.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
include/libxfs.h | 1 -
libxfs/libxfs_io.h | 2 +-
libxfs/rdwr.c | 2 +-
3 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/include/libxfs.h b/include/libxfs.h
index aaac00f6..14113e56 100644
--- a/include/libxfs.h
+++ b/include/libxfs.h
@@ -126,7 +126,6 @@ typedef struct libxfs_xinit {
int bcache_flags; /* cache init flags */
} libxfs_init_t;
-#define LIBXFS_EXIT_ON_FAILURE 0x0001 /* exit the program if a call fails */
#define LIBXFS_ISREADONLY 0x0002 /* disallow all mounted filesystems */
#define LIBXFS_ISINACTIVE 0x0004 /* allow mounted only if mounted ro */
#define LIBXFS_DANGEROUSLY 0x0008 /* repairing a device mounted ro */
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index bb6b689e..560cf0fd 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -86,7 +86,7 @@ bool xfs_verify_magic(struct xfs_buf *bp, __be32 dmagic);
bool xfs_verify_magic16(struct xfs_buf *bp, __be16 dmagic);
/* b_flags bits */
-#define LIBXFS_B_EXIT 0x0001 /* ==LIBXFS_EXIT_ON_FAILURE */
+#define LIBXFS_B_EXIT 0x0001 /* exit if write fails */
#define LIBXFS_B_DIRTY 0x0002 /* buffer has been modified */
#define LIBXFS_B_STALE 0x0004 /* buffer marked as invalid */
#define LIBXFS_B_UPTODATE 0x0008 /* buffer is sync'd to disk */
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index f56303e2..6a75fb12 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -19,7 +19,7 @@
#include "xfs_trans.h"
#include "libfrog/platform.h"
-#include "libxfs.h" /* for LIBXFS_EXIT_ON_FAILURE */
+#include "libxfs.h"
/*
* Important design/architecture note:
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (2 preceding siblings ...)
2020-02-20 1:42 ` [PATCH 03/18] libxfs: remove LIBXFS_EXIT_ON_FAILURE Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:43 ` Christoph Hellwig
2020-02-21 15:01 ` Eric Sandeen
2020-02-20 1:43 ` [PATCH 05/18] libxfs: replace libxfs_getbuf with libxfs_buf_get Darrick J. Wong
` (13 subsequent siblings)
17 siblings, 2 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Change all the libxfs_putbuf calls to libxfs_buf_relse to match the
kernel interface, since one is a #define of the other.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
copy/xfs_copy.c | 4 ++--
db/fsmap.c | 6 +++---
db/info.c | 2 +-
db/init.c | 2 +-
db/io.c | 2 +-
libxfs/init.c | 8 ++++----
libxfs/libxfs_api_defs.h | 1 +
libxfs/libxfs_io.h | 4 ++--
libxfs/libxfs_priv.h | 1 -
libxfs/rdwr.c | 14 +++++++-------
libxfs/trans.c | 10 +++++-----
repair/attr_repair.c | 24 ++++++++++++------------
repair/da_util.c | 22 +++++++++++-----------
repair/dino_chunks.c | 8 ++++----
repair/dinode.c | 8 ++++----
repair/dir2.c | 14 +++++++-------
repair/phase3.c | 2 +-
repair/phase6.c | 36 ++++++++++++++++++------------------
repair/prefetch.c | 10 +++++-----
repair/rmap.c | 8 ++++----
repair/rt.c | 4 ++--
repair/scan.c | 18 +++++++++---------
repair/xfs_repair.c | 2 +-
23 files changed, 105 insertions(+), 105 deletions(-)
diff --git a/copy/xfs_copy.c b/copy/xfs_copy.c
index a6d67038..74dd2e23 100644
--- a/copy/xfs_copy.c
+++ b/copy/xfs_copy.c
@@ -716,12 +716,12 @@ main(int argc, char **argv)
libxfs_sb_from_disk(sb, XFS_BUF_TO_SBP(sbp));
/* Do it again, now with proper length and verifier */
- libxfs_putbuf(sbp);
+ libxfs_buf_relse(sbp);
libxfs_purgebuf(sbp);
sbp = libxfs_readbuf(mbuf.m_ddev_targp, XFS_SB_DADDR,
1 << (sb->sb_sectlog - BBSHIFT),
0, &xfs_sb_buf_ops);
- libxfs_putbuf(sbp);
+ libxfs_buf_relse(sbp);
mp = libxfs_mount(&mbuf, sb, xargs.ddev, xargs.logdev, xargs.rtdev, 0);
if (mp == NULL) {
diff --git a/db/fsmap.c b/db/fsmap.c
index 29f3827c..a6e61962 100644
--- a/db/fsmap.c
+++ b/db/fsmap.c
@@ -75,7 +75,7 @@ fsmap(
bt_cur = libxfs_rmapbt_init_cursor(mp, NULL, agbp, agno);
if (!bt_cur) {
- libxfs_putbuf(agbp);
+ libxfs_buf_relse(agbp);
dbprintf(_("Not enough memory.\n"));
return;
}
@@ -85,14 +85,14 @@ fsmap(
fsmap_fn, &info);
if (error) {
libxfs_btree_del_cursor(bt_cur, XFS_BTREE_ERROR);
- libxfs_putbuf(agbp);
+ libxfs_buf_relse(agbp);
dbprintf(_("Error %d while querying fsmap btree.\n"),
error);
return;
}
libxfs_btree_del_cursor(bt_cur, XFS_BTREE_NOERROR);
- libxfs_putbuf(agbp);
+ libxfs_buf_relse(agbp);
if (agno == start_ag)
low.rm_startblock = 0;
diff --git a/db/info.c b/db/info.c
index fc5ccfe7..5c941dc4 100644
--- a/db/info.c
+++ b/db/info.c
@@ -89,7 +89,7 @@ print_agresv_info(
length = be32_to_cpu(agf->agf_length);
free = be32_to_cpu(agf->agf_freeblks) +
be32_to_cpu(agf->agf_flcount);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
printf("AG %d: length: %u free: %u reserved: %u used: %u",
agno, length, free, ask, used);
diff --git a/db/init.c b/db/init.c
index e92de232..be4a08e5 100644
--- a/db/init.c
+++ b/db/init.c
@@ -123,7 +123,7 @@ init(
/* copy SB from buffer to in-core, converting architecture as we go */
libxfs_sb_from_disk(&xmount.m_sb, XFS_BUF_TO_SBP(bp));
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
libxfs_purgebuf(bp);
sbp = &xmount.m_sb;
diff --git a/db/io.c b/db/io.c
index 6b66472e..a11b7bb1 100644
--- a/db/io.c
+++ b/db/io.c
@@ -96,7 +96,7 @@ pop_cur(void)
return;
}
if (iocur_top->bp) {
- libxfs_putbuf(iocur_top->bp);
+ libxfs_buf_relse(iocur_top->bp);
iocur_top->bp = NULL;
}
if (iocur_top->bbmap) {
diff --git a/libxfs/init.c b/libxfs/init.c
index 428497f0..2c290428 100644
--- a/libxfs/init.c
+++ b/libxfs/init.c
@@ -445,7 +445,7 @@ rtmount_init(
progname);
return -1;
}
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 0;
}
@@ -729,7 +729,7 @@ libxfs_mount(
if (!(flags & LIBXFS_MOUNT_DEBUGGER))
return NULL;
} else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
if (mp->m_logdev_targp->dev &&
mp->m_logdev_targp->dev != mp->m_ddev_targp->dev) {
@@ -745,7 +745,7 @@ libxfs_mount(
return NULL;
}
if (bp)
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
/* Initialize realtime fields in the mount structure */
@@ -775,7 +775,7 @@ libxfs_mount(
progname);
sbp->sb_agcount = 1;
}
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
error = libxfs_initialize_perag(mp, sbp->sb_agcount, &mp->m_maxagi);
diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 1a438e58..8bca7ddf 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -44,6 +44,7 @@
#define xfs_btree_del_cursor libxfs_btree_del_cursor
#define xfs_btree_init_block libxfs_btree_init_block
#define xfs_buf_delwri_submit libxfs_buf_delwri_submit
+#define xfs_buf_relse libxfs_buf_relse
#define xfs_bunmapi libxfs_bunmapi
#define xfs_calc_dquots_per_chunk libxfs_calc_dquots_per_chunk
#define xfs_da3_node_hdr_from_disk libxfs_da3_node_hdr_from_disk
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 560cf0fd..704237d1 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -150,7 +150,7 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_getbuf_flags(dev, daddr, len, flags) \
libxfs_trace_getbuf_flags(__FUNCTION__, __FILE__, __LINE__, \
(dev), (daddr), (len), (flags))
-#define libxfs_putbuf(buf) \
+#define libxfs_buf_relse(buf) \
libxfs_trace_putbuf(__FUNCTION__, __FILE__, __LINE__, (buf))
extern xfs_buf_t *libxfs_trace_readbuf(const char *, const char *, int,
@@ -182,7 +182,7 @@ extern xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *,
struct xfs_buf_map *, int, int);
extern xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t,
int, unsigned int);
-extern void libxfs_putbuf (xfs_buf_t *);
+void libxfs_buf_relse(struct xfs_buf *);
#endif
diff --git a/libxfs/libxfs_priv.h b/libxfs/libxfs_priv.h
index fe08f96b..4bd3c462 100644
--- a/libxfs/libxfs_priv.h
+++ b/libxfs/libxfs_priv.h
@@ -377,7 +377,6 @@ roundup_64(uint64_t x, uint32_t y)
(len) = __bar; /* no set-but-unused warning */ \
NULL; \
})
-#define xfs_buf_relse(bp) libxfs_putbuf(bp)
#define xfs_buf_get(devp,blkno,len) (libxfs_getbuf((devp), (blkno), (len)))
#define xfs_bwrite(bp) libxfs_writebuf((bp), 0)
#define xfs_buf_oneshot(bp) ((void) 0)
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 6a75fb12..62b79bc4 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -48,7 +48,7 @@
*
* The result of this is that until the userspace code outside libxfs is cleaned
* up, functions that release buffers from userspace control (i.e
- * libxfs_writebuf/libxfs_putbuf) need to zero bp->b_error to prevent
+ * libxfs_writebuf/libxfs_buf_relse) need to zero bp->b_error to prevent
* propagation of stale errors into future buffer operations.
*/
@@ -387,7 +387,6 @@ libxfs_log_header(
#undef libxfs_getbuf
#undef libxfs_getbuf_map
#undef libxfs_getbuf_flags
-#undef libxfs_putbuf
xfs_buf_t *libxfs_readbuf(struct xfs_buftarg *, xfs_daddr_t, int, int,
const struct xfs_buf_ops *);
@@ -399,7 +398,7 @@ xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int);
xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t, int,
unsigned int);
-void libxfs_putbuf (xfs_buf_t *);
+void libxfs_buf_relse(struct xfs_buf *);
#define __add_trace(bp, func, file, line) \
do { \
@@ -469,7 +468,7 @@ void
libxfs_trace_putbuf(const char *func, const char *file, int line, xfs_buf_t *bp)
{
__add_trace(bp, func, file, line);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
@@ -844,7 +843,8 @@ libxfs_getbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
}
void
-libxfs_putbuf(xfs_buf_t *bp)
+libxfs_buf_relse(
+ struct xfs_buf *bp)
{
/*
* ensure that any errors on this use of the buffer don't carry
@@ -1199,7 +1199,7 @@ libxfs_writebuf(
bp->b_error = 0;
bp->b_flags &= ~LIBXFS_B_STALE;
bp->b_flags |= bflags;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 0;
}
@@ -1529,7 +1529,7 @@ xfs_buf_delwri_submit(
error2 = libxfs_writebufr(bp);
if (!error)
error = error2;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
return error;
diff --git a/libxfs/trans.c b/libxfs/trans.c
index 18b87d70..59cb897f 100644
--- a/libxfs/trans.c
+++ b/libxfs/trans.c
@@ -566,7 +566,7 @@ libxfs_trans_brelse(
ASSERT(bp->b_transp == tp);
if (!tp) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return;
}
@@ -602,7 +602,7 @@ libxfs_trans_brelse(
xfs_buf_item_put(bip);
bp->b_transp = NULL;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
/*
@@ -839,7 +839,7 @@ inode_item_done(
if (error) {
fprintf(stderr, _("%s: warning - iflush_int failed (%d)\n"),
progname, error);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto free;
}
@@ -868,7 +868,7 @@ buf_item_done(
xfs_buf_item_put(bip);
if (hold)
return;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
static void
@@ -906,7 +906,7 @@ buf_item_unlock(
bip->bli_flags &= ~XFS_BLI_HOLD;
xfs_buf_item_put(bip);
if (!hold)
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
static void
diff --git a/repair/attr_repair.c b/repair/attr_repair.c
index 535fcfbb..ec068ba2 100644
--- a/repair/attr_repair.c
+++ b/repair/attr_repair.c
@@ -418,7 +418,7 @@ rmtval_get(xfs_mount_t *mp, xfs_ino_t ino, blkmap_t *blkmap,
if (bp->b_error == -EFSBADCRC || bp->b_error == -EFSCORRUPTED) {
do_warn(
_("Corrupt remote block for attributes of inode %" PRIu64 "\n"), ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
clearit = 1;
break;
}
@@ -430,7 +430,7 @@ rmtval_get(xfs_mount_t *mp, xfs_ino_t ino, blkmap_t *blkmap,
amountdone += length;
value += length;
i++;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
return (clearit);
}
@@ -782,7 +782,7 @@ process_leaf_attr_level(xfs_mount_t *mp,
do_warn(
_("bad attribute leaf magic %#x for inode %" PRIu64 "\n"),
leafhdr.magic, ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
@@ -793,7 +793,7 @@ process_leaf_attr_level(xfs_mount_t *mp,
if (process_leaf_attr_block(mp, leaf, da_bno, ino,
da_cursor->blkmap, current_hashval,
&greatest_hashval, &repair)) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
@@ -813,7 +813,7 @@ process_leaf_attr_level(xfs_mount_t *mp,
do_warn(
_("bad sibling back pointer for block %u in attribute fork for inode %" PRIu64 "\n"),
da_bno, ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
@@ -822,7 +822,7 @@ process_leaf_attr_level(xfs_mount_t *mp,
if (da_bno != 0) {
if (verify_da_path(mp, da_cursor, 0, XFS_ATTR_FORK)) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
}
@@ -838,7 +838,7 @@ process_leaf_attr_level(xfs_mount_t *mp,
if (repair && !no_modify)
libxfs_writebuf(bp, 0);
else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
} while (da_bno != 0);
if (verify_final_da_path(mp, da_cursor, 0, XFS_ATTR_FORK)) {
@@ -992,7 +992,7 @@ _("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "
if (badness) {
*repair = 0;
/* the block is bad. lose the attribute fork. */
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
@@ -1001,7 +1001,7 @@ _("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "
if (*repair && !no_modify)
libxfs_writebuf(bp, 0);
else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 0;
}
@@ -1045,7 +1045,7 @@ _("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "
*repair = 1;
libxfs_writebuf(bp, 0);
} else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
error = process_node_attr(mp, ino, dip, blkmap); /* + repair */
if (error)
*repair = 0;
@@ -1107,7 +1107,7 @@ process_longform_attr(
/* is this block sane? */
if (__check_attr_header(mp, bp, ino)) {
*repair = 0;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
@@ -1130,7 +1130,7 @@ process_longform_attr(
do_warn(
_("bad attribute leaf magic # %#x for dir ino %" PRIu64 "\n"),
be16_to_cpu(info->magic), ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
*repair = 0;
return 1;
}
diff --git a/repair/da_util.c b/repair/da_util.c
index bf25f7f2..c02d621c 100644
--- a/repair/da_util.c
+++ b/repair/da_util.c
@@ -145,7 +145,7 @@ _("found non-root LEAFN node in inode %" PRIu64 " bno = %u\n"),
da_cursor->ino, bno);
}
*rbno = 0;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
@@ -155,13 +155,13 @@ _("found non-root LEAFN node in inode %" PRIu64 " bno = %u\n"),
_("bad %s magic number 0x%x in inode %" PRIu64 " bno = %u\n"),
FORKNAME(whichfork), nodehdr.magic,
da_cursor->ino, bno);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
/* corrupt node; rebuild the dir. */
if (bp->b_error == -EFSBADCRC || bp->b_error == -EFSCORRUPTED) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
do_warn(
_("corrupt %s tree block %u for inode %" PRIu64 "\n"),
FORKNAME(whichfork), bno, da_cursor->ino);
@@ -173,7 +173,7 @@ _("corrupt %s tree block %u for inode %" PRIu64 "\n"),
_("bad %s record count in inode %" PRIu64 ", count = %d, max = %d\n"),
FORKNAME(whichfork), da_cursor->ino,
nodehdr.count, geo->node_ents);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
@@ -186,7 +186,7 @@ _("bad %s record count in inode %" PRIu64 ", count = %d, max = %d\n"),
do_warn(
_("bad header depth for directory inode %" PRIu64 "\n"),
da_cursor->ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
i = -1;
goto error_out;
}
@@ -197,7 +197,7 @@ _("bad header depth for directory inode %" PRIu64 "\n"),
do_warn(
_("bad %s btree for inode %" PRIu64 "\n"),
FORKNAME(whichfork), da_cursor->ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
}
@@ -222,7 +222,7 @@ _("bad %s btree for inode %" PRIu64 "\n"),
error_out:
while (i > 1 && i <= da_cursor->active) {
- libxfs_putbuf(da_cursor->level[i].bp);
+ libxfs_buf_relse(da_cursor->level[i].bp);
i++;
}
@@ -252,7 +252,7 @@ release_da_cursor_int(
}
ASSERT(error != 0);
- libxfs_putbuf(cursor->level[level].bp);
+ libxfs_buf_relse(cursor->level[level].bp);
cursor->level[level].bp = NULL;
}
@@ -406,7 +406,7 @@ _("would correct bad hashval in non-leaf %s block\n"
if (cursor->level[this_level].dirty && !no_modify)
libxfs_writebuf(cursor->level[this_level].bp, 0);
else
- libxfs_putbuf(cursor->level[this_level].bp);
+ libxfs_buf_relse(cursor->level[this_level].bp);
cursor->level[this_level].bp = NULL;
@@ -600,7 +600,7 @@ _("bad level %d in %s block %u for inode %" PRIu64 "\n"),
#ifdef XR_DIR_TRACE
fprintf(stderr, "verify_da_path returns 1 (bad) #4\n");
#endif
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
@@ -622,7 +622,7 @@ _("bad level %d in %s block %u for inode %" PRIu64 "\n"),
if (cursor->level[this_level].dirty && !no_modify)
libxfs_writebuf(cursor->level[this_level].bp, 0);
else
- libxfs_putbuf(cursor->level[this_level].bp);
+ libxfs_buf_relse(cursor->level[this_level].bp);
/* switch cursor to point at the new buffer we just read */
cursor->level[this_level].bp = bp;
diff --git a/repair/dino_chunks.c b/repair/dino_chunks.c
index dbf3d37a..863c4531 100644
--- a/repair/dino_chunks.c
+++ b/repair/dino_chunks.c
@@ -56,7 +56,7 @@ check_aginode_block(xfs_mount_t *mp,
if (cnt)
bp->b_ops = &xfs_inode_buf_ops;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return(cnt);
}
@@ -670,7 +670,7 @@ process_inode_chunk(
M_IGEO(mp)->blocks_per_cluster));
while (bp_index > 0) {
bp_index--;
- libxfs_putbuf(bplist[bp_index]);
+ libxfs_buf_relse(bplist[bp_index]);
}
free(bplist);
return(1);
@@ -759,7 +759,7 @@ process_inode_chunk(
*bogus = 1;
for (bp_index = 0; bp_index < cluster_count; bp_index++)
if (bplist[bp_index])
- libxfs_putbuf(bplist[bp_index]);
+ libxfs_buf_relse(bplist[bp_index]);
free(bplist);
return(0);
}
@@ -942,7 +942,7 @@ process_inode_chunk(
if (dirty && !no_modify)
libxfs_writebuf(bplist[bp_index], 0);
else
- libxfs_putbuf(bplist[bp_index]);
+ libxfs_buf_relse(bplist[bp_index]);
}
free(bplist);
break;
diff --git a/repair/dinode.c b/repair/dinode.c
index 0d9c96be..928698cb 100644
--- a/repair/dinode.c
+++ b/repair/dinode.c
@@ -1234,7 +1234,7 @@ _("cannot read inode %" PRIu64 ", file block %" PRIu64 ", disk block %" PRIu64 "
if (writebuf && !no_modify)
libxfs_writebuf(bp, 0);
else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
return 0;
}
@@ -1296,7 +1296,7 @@ _("cannot read inode %" PRIu64 ", file block %d, disk block %" PRIu64 "\n"),
do_warn(
_("Corrupt symlink remote block %" PRIu64 ", inode %" PRIu64 ".\n"),
fsbno, lino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
if (bp->b_error == -EFSBADCRC) {
@@ -1316,7 +1316,7 @@ _("Bad symlink buffer CRC, block %" PRIu64 ", inode %" PRIu64 ".\n"
do_warn(
_("bad symlink header ino %" PRIu64 ", file block %d, disk block %" PRIu64 "\n"),
lino, i, fsbno);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
src += sizeof(struct xfs_dsymlink_hdr);
@@ -1331,7 +1331,7 @@ _("bad symlink header ino %" PRIu64 ", file block %d, disk block %" PRIu64 "\n")
if (badcrc && !no_modify)
libxfs_writebuf(bp, 0);
else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
return 0;
}
diff --git a/repair/dir2.c b/repair/dir2.c
index 723aee1f..769e341c 100644
--- a/repair/dir2.c
+++ b/repair/dir2.c
@@ -1012,7 +1012,7 @@ _("bad directory block magic # %#x in block %u for directory inode %" PRIu64 "\n
*repair = 1;
libxfs_writebuf(bp, 0);
} else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return rval;
}
@@ -1131,7 +1131,7 @@ _("can't read file block %u for directory inode %" PRIu64 "\n"),
do_warn(
_("bad directory leaf magic # %#x for directory inode %" PRIu64 " block %u\n"),
leafhdr.magic, ino, da_bno);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
buf_dirty = 0;
@@ -1141,7 +1141,7 @@ _("bad directory leaf magic # %#x for directory inode %" PRIu64 " block %u\n"),
*/
if (process_leaf_block_dir2(mp, leaf, da_bno, ino,
current_hashval, &greatest_hashval)) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
/*
@@ -1159,14 +1159,14 @@ _("bad directory leaf magic # %#x for directory inode %" PRIu64 " block %u\n"),
do_warn(
_("bad sibling back pointer for block %u in directory inode %" PRIu64 "\n"),
da_bno, ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
prev_bno = da_bno;
da_bno = leafhdr.forw;
if (da_bno != 0) {
if (verify_da_path(mp, da_cursor, 0, XFS_DATA_FORK)) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
goto error_out;
}
}
@@ -1182,7 +1182,7 @@ _("bad sibling back pointer for block %u in directory inode %" PRIu64 "\n"),
*repair = 1;
libxfs_writebuf(bp, 0);
} else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
} while (da_bno != 0);
if (verify_final_da_path(mp, da_cursor, 0, XFS_DATA_FORK)) {
/*
@@ -1341,7 +1341,7 @@ _("bad directory block magic # %#x in block %" PRIu64 " for directory inode %" P
*repair = 1;
libxfs_writebuf(bp, 0);
} else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
if (good == 0)
return 1;
diff --git a/repair/phase3.c b/repair/phase3.c
index 161852e0..1c6929ac 100644
--- a/repair/phase3.c
+++ b/repair/phase3.c
@@ -49,7 +49,7 @@ process_agi_unlinked(
if (agi_dirty)
libxfs_writebuf(bp, 0);
else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
static void
diff --git a/repair/phase6.c b/repair/phase6.c
index 70135694..2464fe1a 100644
--- a/repair/phase6.c
+++ b/repair/phase6.c
@@ -1624,7 +1624,7 @@ longform_dir2_entry_check_data(
dir2_kill_block(mp, ip, da_bno, bp);
} else {
do_warn(_("would junk block\n"));
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
freetab->ents[db].v = NULLDATAOFF;
*bpp = NULL;
@@ -2063,21 +2063,21 @@ longform_dir2_check_leaf(
do_warn(
_("leaf block %u for directory inode %" PRIu64 " bad header\n"),
da_bno, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
if (leafhdr.magic == XFS_DIR3_LEAF1_MAGIC) {
error = check_da3_header(mp, bp, ip->i_ino);
if (error) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return error;
}
}
seeval = dir_hash_see_all(hashtab, ents, leafhdr.count, leafhdr.stale);
if (dir_hash_check(hashtab, ip, seeval)) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
badtail = freetab->nents != be32_to_cpu(ltp->bestcount);
@@ -2089,10 +2089,10 @@ longform_dir2_check_leaf(
do_warn(
_("leaf block %u for directory inode %" PRIu64 " bad tail\n"),
da_bno, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return fixit;
}
@@ -2155,7 +2155,7 @@ longform_dir2_check_node(
do_warn(
_("unknown magic number %#x for block %u in directory inode %" PRIu64 "\n"),
leafhdr.magic, da_bno, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
@@ -2164,7 +2164,7 @@ longform_dir2_check_node(
leafhdr.magic == XFS_DA3_NODE_MAGIC) {
error = check_da3_header(mp, bp, ip->i_ino);
if (error) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return error;
}
}
@@ -2172,7 +2172,7 @@ longform_dir2_check_node(
/* ignore nodes */
if (leafhdr.magic == XFS_DA_NODE_MAGIC ||
leafhdr.magic == XFS_DA3_NODE_MAGIC) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
continue;
}
@@ -2186,12 +2186,12 @@ longform_dir2_check_node(
do_warn(
_("leaf block %u for directory inode %" PRIu64 " bad header\n"),
da_bno, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
seeval = dir_hash_see_all(hashtab, ents,
leafhdr.count, leafhdr.stale);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
if (seeval != DIR_HASH_CK_OK)
return 1;
}
@@ -2226,14 +2226,14 @@ longform_dir2_check_node(
do_warn(
_("free block %u for directory inode %" PRIu64 " bad header\n"),
da_bno, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
if (freehdr.magic == XFS_DIR3_FREE_MAGIC) {
error = check_dir3_header(mp, bp, ip->i_ino);
if (error) {
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return error;
}
}
@@ -2244,7 +2244,7 @@ longform_dir2_check_node(
do_warn(
_("free block %u entry %i for directory ino %" PRIu64 " bad\n"),
da_bno, i, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
used += be16_to_cpu(bests[i]) != NULLDATAOFF;
@@ -2254,10 +2254,10 @@ longform_dir2_check_node(
do_warn(
_("free block %u for directory inode %" PRIu64 " bad nused\n"),
da_bno, ip->i_ino);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 1;
}
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
for (i = 0; i < freetab->nents; i++) {
if ((freetab->ents[i].s == 0) &&
@@ -2433,14 +2433,14 @@ longform_dir2_entry_check(xfs_mount_t *mp,
dir_hash_dup_names(hashtab);
for (i = 0; i < num_bps; i++)
if (bplist[i])
- libxfs_putbuf(bplist[i]);
+ libxfs_buf_relse(bplist[i]);
longform_dir2_rebuild(mp, ino, ip, irec, ino_offset, hashtab);
*num_illegal = 0;
*need_dot = 0;
} else {
for (i = 0; i < num_bps; i++)
if (bplist[i])
- libxfs_putbuf(bplist[i]);
+ libxfs_buf_relse(bplist[i]);
}
free(bplist);
diff --git a/repair/prefetch.c b/repair/prefetch.c
index 8e3772ed..f7ea9c8f 100644
--- a/repair/prefetch.c
+++ b/repair/prefetch.c
@@ -129,7 +129,7 @@ pf_queue_io(
pf_read_inode_dirs(args, bp);
XFS_BUF_SET_PRIORITY(bp, XFS_BUF_PRIORITY(bp) +
CACHE_PREFETCH_PRIORITY);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return;
}
XFS_BUF_SET_PRIORITY(bp, flag);
@@ -286,13 +286,13 @@ pf_scan_lbtree(
*/
if (bp->b_error) {
bp->b_flags |= LIBXFS_B_UNCHECKED;
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return 0;
}
rc = (*func)(XFS_BUF_TO_BLOCK(bp), level - 1, isadir, args);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return rc;
}
@@ -578,7 +578,7 @@ pf_batch_read(
if ((bplist[num - 1]->b_flags & LIBXFS_B_DISCONTIG)) {
libxfs_readbufr_map(mp->m_ddev_targp, bplist[num - 1], 0);
bplist[num - 1]->b_flags |= LIBXFS_B_UNCHECKED;
- libxfs_putbuf(bplist[num - 1]);
+ libxfs_buf_relse(bplist[num - 1]);
num--;
}
@@ -612,7 +612,7 @@ pf_batch_read(
B_IS_INODE(XFS_BUF_PRIORITY(bplist[i])) ? 'I' : 'M',
bplist[i], (long long)XFS_BUF_ADDR(bplist[i]),
args->agno);
- libxfs_putbuf(bplist[i]);
+ libxfs_buf_relse(bplist[i]);
}
pthread_mutex_lock(&args->lock);
if (which != PF_SECONDARY) {
diff --git a/repair/rmap.c b/repair/rmap.c
index 2f99d35d..bc53e6c0 100644
--- a/repair/rmap.c
+++ b/repair/rmap.c
@@ -527,7 +527,7 @@ rmap_store_ag_btree_rec(
}
b++;
}
- libxfs_putbuf(agflbp);
+ libxfs_buf_relse(agflbp);
agflbp = NULL;
bitmap_free(&own_ag_bitmap);
@@ -579,7 +579,7 @@ rmap_store_ag_btree_rec(
free_slab_cursor(&rm_cur);
err:
if (agflbp)
- libxfs_putbuf(agflbp);
+ libxfs_buf_relse(agflbp);
if (own_ag_bitmap)
bitmap_free(&own_ag_bitmap);
return error;
@@ -1082,7 +1082,7 @@ _("Incorrect reverse-mapping: saw (%u/%u) %slen %u owner %"PRId64" %s%soff \
if (bt_cur)
libxfs_btree_del_cursor(bt_cur, XFS_BTREE_NOERROR);
if (agbp)
- libxfs_putbuf(agbp);
+ libxfs_buf_relse(agbp);
free_slab_cursor(&rm_cur);
return 0;
}
@@ -1417,7 +1417,7 @@ _("Incorrect reference count: saw (%u/%u) len %u nlinks %u; should be (%u/%u) le
libxfs_btree_del_cursor(bt_cur, error ? XFS_BTREE_ERROR :
XFS_BTREE_NOERROR);
if (agbp)
- libxfs_putbuf(agbp);
+ libxfs_buf_relse(agbp);
free_slab_cursor(&rl_cur);
return 0;
}
diff --git a/repair/rt.c b/repair/rt.c
index 7108e3d5..3319829c 100644
--- a/repair/rt.c
+++ b/repair/rt.c
@@ -222,7 +222,7 @@ process_rtbitmap(xfs_mount_t *mp,
prevbit = 0;
}
}
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
if (extno == mp->m_sb.sb_rextents)
break;
}
@@ -266,7 +266,7 @@ process_rtsummary(xfs_mount_t *mp,
bytes = bp->b_un.b_addr;
memmove((char *)sumfile + sumbno * mp->m_sb.sb_blocksize, bytes,
mp->m_sb.sb_blocksize);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
}
#endif
diff --git a/repair/scan.c b/repair/scan.c
index 05707dd2..c961e843 100644
--- a/repair/scan.c
+++ b/repair/scan.c
@@ -82,7 +82,7 @@ scan_sbtree(
(*func)(XFS_BUF_TO_BLOCK(bp), nlevels - 1, root, agno, suspect,
isroot, magic, priv, ops);
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
}
/*
@@ -154,7 +154,7 @@ scan_lbtree(
if ((dirty || badcrc) && !no_modify)
libxfs_writebuf(bp, 0);
else
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
return(err);
}
@@ -2142,7 +2142,7 @@ scan_freelist(
agcnts->fdblocks += state.count;
- libxfs_putbuf(agflbuf);
+ libxfs_buf_relse(agflbuf);
}
static void
@@ -2432,12 +2432,12 @@ scan_ag(
if (agi_dirty && !no_modify)
libxfs_writebuf(agibuf, 0);
else
- libxfs_putbuf(agibuf);
+ libxfs_buf_relse(agibuf);
if (agf_dirty && !no_modify)
libxfs_writebuf(agfbuf, 0);
else
- libxfs_putbuf(agfbuf);
+ libxfs_buf_relse(agfbuf);
if (sb_dirty && !no_modify) {
if (agno == 0)
@@ -2445,7 +2445,7 @@ scan_ag(
libxfs_sb_to_disk(XFS_BUF_TO_SBP(sbbuf), sb);
libxfs_writebuf(sbbuf, 0);
} else
- libxfs_putbuf(sbbuf);
+ libxfs_buf_relse(sbbuf);
free(sb);
PROG_RPT_INC(prog_rpt_done[agno], 1);
@@ -2455,11 +2455,11 @@ scan_ag(
return;
out_free_agibuf:
- libxfs_putbuf(agibuf);
+ libxfs_buf_relse(agibuf);
out_free_agfbuf:
- libxfs_putbuf(agfbuf);
+ libxfs_buf_relse(agfbuf);
out_free_sbbuf:
- libxfs_putbuf(sbbuf);
+ libxfs_buf_relse(sbbuf);
out_free_sb:
free(sb);
diff --git a/repair/xfs_repair.c b/repair/xfs_repair.c
index c0a77cad..8642c5cd 100644
--- a/repair/xfs_repair.c
+++ b/repair/xfs_repair.c
@@ -479,7 +479,7 @@ guess_correct_sunit(
if (error)
continue;
libxfs_sb_from_disk(&sb, XFS_BUF_TO_SBP(bp));
- libxfs_putbuf(bp);
+ libxfs_buf_relse(bp);
calc_rootino = libxfs_ialloc_calc_rootino(mp, sb.sb_unit);
if (calc_rootino == mp->m_sb.sb_rootino)
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 05/18] libxfs: replace libxfs_getbuf with libxfs_buf_get
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (3 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:43 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 06/18] libxfs: replace libxfs_readbuf with libxfs_buf_read Darrick J. Wong
` (12 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Change all the libxfs_getbuf calls to libxfs_buf_get to match the
kernel interface, since one is a #define of the other.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
include/cache.h | 2 +-
libxfs/libxfs_api_defs.h | 1 +
libxfs/libxfs_io.h | 21 ++++++++++++++++-----
libxfs/libxfs_priv.h | 1 -
libxfs/rdwr.c | 34 ++++++++++++++++------------------
mkfs/xfs_mkfs.c | 8 ++++----
repair/phase5.c | 30 +++++++++++++++---------------
7 files changed, 53 insertions(+), 44 deletions(-)
diff --git a/include/cache.h b/include/cache.h
index 552b9248..334ad263 100644
--- a/include/cache.h
+++ b/include/cache.h
@@ -33,7 +33,7 @@ enum {
* For prefetch support, the top half of the range starts at
* CACHE_PREFETCH_PRIORITY and everytime the buffer is fetched and is at or
* above this priority level, it is reduced to below this level (refer to
- * libxfs_getbuf).
+ * libxfs_buf_get).
*
* If we have dirty nodes, we can't recycle them until they've been cleaned. To
* keep these out of the reclaimable lists (as there can be lots of them) give
diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 8bca7ddf..6d86774f 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -44,6 +44,7 @@
#define xfs_btree_del_cursor libxfs_btree_del_cursor
#define xfs_btree_init_block libxfs_btree_init_block
#define xfs_buf_delwri_submit libxfs_buf_delwri_submit
+#define xfs_buf_get libxfs_buf_get
#define xfs_buf_relse libxfs_buf_relse
#define xfs_bunmapi libxfs_bunmapi
#define xfs_calc_dquots_per_chunk libxfs_calc_dquots_per_chunk
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 704237d1..5087f03d 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -141,7 +141,7 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_writebuf(buf, flags) \
libxfs_trace_writebuf(__FUNCTION__, __FILE__, __LINE__, \
(buf), (flags))
-#define libxfs_getbuf(dev, daddr, len) \
+#define libxfs_buf_get(dev, daddr, len) \
libxfs_trace_getbuf(__FUNCTION__, __FILE__, __LINE__, \
(dev), (daddr), (len))
#define libxfs_getbuf_map(dev, map, nmaps, flags) \
@@ -161,8 +161,9 @@ extern xfs_buf_t *libxfs_trace_readbuf_map(const char *, const char *, int,
const struct xfs_buf_ops *);
extern int libxfs_trace_writebuf(const char *, const char *, int,
xfs_buf_t *, int);
-extern xfs_buf_t *libxfs_trace_getbuf(const char *, const char *, int,
- struct xfs_buftarg *, xfs_daddr_t, int);
+struct xfs_buf *libxfs_trace_getbuf(const char *func, const char *file,
+ int line, struct xfs_buftarg *btp, xfs_daddr_t daddr,
+ size_t len);
extern xfs_buf_t *libxfs_trace_getbuf_map(const char *, const char *, int,
struct xfs_buftarg *, struct xfs_buf_map *, int, int);
extern xfs_buf_t *libxfs_trace_getbuf_flags(const char *, const char *, int,
@@ -177,14 +178,24 @@ extern xfs_buf_t *libxfs_readbuf(struct xfs_buftarg *, xfs_daddr_t, int, int,
extern xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
extern int libxfs_writebuf(xfs_buf_t *, int);
-extern xfs_buf_t *libxfs_getbuf(struct xfs_buftarg *, xfs_daddr_t, int);
extern xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *,
struct xfs_buf_map *, int, int);
extern xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t,
int, unsigned int);
void libxfs_buf_relse(struct xfs_buf *);
-#endif
+static inline struct xfs_buf*
+libxfs_buf_get(
+ struct xfs_buftarg *target,
+ xfs_daddr_t blkno,
+ size_t numblks)
+{
+ DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+
+ return libxfs_getbuf_map(target, &map, 1, 0);
+}
+
+#endif /* XFS_BUF_TRACING */
extern void libxfs_readbuf_verify(struct xfs_buf *bp,
const struct xfs_buf_ops *ops);
diff --git a/libxfs/libxfs_priv.h b/libxfs/libxfs_priv.h
index 4bd3c462..0f26120f 100644
--- a/libxfs/libxfs_priv.h
+++ b/libxfs/libxfs_priv.h
@@ -377,7 +377,6 @@ roundup_64(uint64_t x, uint32_t y)
(len) = __bar; /* no set-but-unused warning */ \
NULL; \
})
-#define xfs_buf_get(devp,blkno,len) (libxfs_getbuf((devp), (blkno), (len)))
#define xfs_bwrite(bp) libxfs_writebuf((bp), 0)
#define xfs_buf_oneshot(bp) ((void) 0)
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 62b79bc4..56c50d47 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -30,7 +30,7 @@
* outside libxfs clears bp->b_error - very little code even checks it - so the
* libxfs code is tripping on stale errors left by the userspace code.
*
- * We can't clear errors or zero buffer contents in libxfs_getbuf-* like we do
+ * We can't clear errors or zero buffer contents in libxfs_buf_get-* like we do
* in the kernel, because those functions are used by the libxfs_readbuf_*
* functions and hence need to leave the buffers unchanged on cache hits. This
* is actually the only way to gather a write error from a libxfs_writebuf()
@@ -44,7 +44,7 @@
*
* IOWs, userspace is behaving quite differently to the kernel and as a result
* it leaks errors from reads, invalidations and writes through
- * libxfs_getbuf/libxfs_readbuf.
+ * libxfs_buf_get/libxfs_readbuf.
*
* The result of this is that until the userspace code outside libxfs is cleaned
* up, functions that release buffers from userspace control (i.e
@@ -384,7 +384,6 @@ libxfs_log_header(
#undef libxfs_readbuf
#undef libxfs_readbuf_map
#undef libxfs_writebuf
-#undef libxfs_getbuf
#undef libxfs_getbuf_map
#undef libxfs_getbuf_flags
@@ -393,7 +392,8 @@ xfs_buf_t *libxfs_readbuf(struct xfs_buftarg *, xfs_daddr_t, int, int,
xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
int libxfs_writebuf(xfs_buf_t *, int);
-xfs_buf_t *libxfs_getbuf(struct xfs_buftarg *, xfs_daddr_t, int);
+struct xfs_buf *libxfs_buf_get(struct xfs_buftarg *btp, xfs_daddr_t daddr,
+ size_t len);
xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int);
xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t, int,
@@ -436,11 +436,19 @@ libxfs_trace_writebuf(const char *func, const char *file, int line, xfs_buf_t *b
return libxfs_writebuf(bp, flags);
}
-xfs_buf_t *
-libxfs_trace_getbuf(const char *func, const char *file, int line,
- struct xfs_buftarg *btp, xfs_daddr_t blkno, int len)
+struct xfs_buf *
+libxfs_trace_getbuf(
+ const char *func,
+ const char *file,
+ int line,
+ struct xfs_buftarg *btp,
+ xfs_daddr_t blkno,
+ size_t len)
{
- xfs_buf_t *bp = libxfs_getbuf(btp, blkno, len);
+ struct xfs_buf *bp;
+ DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+
+ bp = libxfs_getbuf_map(target, &map, 1, 0);
__add_trace(bp, func, file, line);
return bp;
}
@@ -799,16 +807,6 @@ reset_buf_state(
LIBXFS_B_UPTODATE);
}
-struct xfs_buf *
-libxfs_getbuf(struct xfs_buftarg *btp, xfs_daddr_t blkno, int len)
-{
- struct xfs_buf *bp;
-
- bp = libxfs_getbuf_flags(btp, blkno, len, 0);
- reset_buf_state(bp);
- return bp;
-}
-
static struct xfs_buf *
__libxfs_getbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
int nmaps, int flags)
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index db640a11..948749e9 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -3444,7 +3444,7 @@ prepare_devices(
* the end of the device. (MD sb is ~64k from the end, take out a wider
* swath to be sure)
*/
- buf = libxfs_getbuf(mp->m_ddev_targp, (xi->dsize - whack_blks),
+ buf = libxfs_buf_get(mp->m_ddev_targp, (xi->dsize - whack_blks),
whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
@@ -3456,13 +3456,13 @@ prepare_devices(
* swap (somewhere around the page size), jfs (32k),
* ext[2,3] and reiserfs (64k) - and hopefully all else.
*/
- buf = libxfs_getbuf(mp->m_ddev_targp, 0, whack_blks);
+ buf = libxfs_buf_get(mp->m_ddev_targp, 0, whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
libxfs_purgebuf(buf);
/* OK, now write the superblock... */
- buf = libxfs_getbuf(mp->m_ddev_targp, XFS_SB_DADDR, XFS_FSS_TO_BB(mp, 1));
+ buf = libxfs_buf_get(mp->m_ddev_targp, XFS_SB_DADDR, XFS_FSS_TO_BB(mp, 1));
buf->b_ops = &xfs_sb_buf_ops;
memset(buf->b_addr, 0, cfg->sectorsize);
libxfs_sb_to_disk(buf->b_addr, sbp);
@@ -3482,7 +3482,7 @@ prepare_devices(
/* finally, check we can write the last block in the realtime area */
if (mp->m_rtdev_targp->dev && cfg->rtblocks > 0) {
- buf = libxfs_getbuf(mp->m_rtdev_targp,
+ buf = libxfs_buf_get(mp->m_rtdev_targp,
XFS_FSB_TO_BB(mp, cfg->rtblocks - 1LL),
BTOBB(cfg->blocksize));
memset(buf->b_addr, 0, cfg->blocksize);
diff --git a/repair/phase5.c b/repair/phase5.c
index 7f7d3d18..cdbf6697 100644
--- a/repair/phase5.c
+++ b/repair/phase5.c
@@ -689,7 +689,7 @@ prop_freespace_cursor(xfs_mount_t *mp, xfs_agnumber_t agno,
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
lptr->agbno = agbno;
@@ -767,7 +767,7 @@ build_freespace_tree(xfs_mount_t *mp, xfs_agnumber_t agno,
lptr = &btree_curs->level[i];
agbno = get_next_blockaddr(agno, i, btree_curs);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
@@ -877,7 +877,7 @@ build_freespace_tree(xfs_mount_t *mp, xfs_agnumber_t agno,
lptr->agbno = get_next_blockaddr(agno, 0, btree_curs);
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(lptr->agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, lptr->agbno),
XFS_FSB_TO_BB(mp, 1));
}
@@ -1054,7 +1054,7 @@ prop_ino_cursor(xfs_mount_t *mp, xfs_agnumber_t agno, bt_status_t *btree_curs,
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
lptr->agbno = agbno;
@@ -1104,7 +1104,7 @@ build_agi(xfs_mount_t *mp, xfs_agnumber_t agno, bt_status_t *btree_curs,
xfs_agi_t *agi;
int i;
- agi_buf = libxfs_getbuf(mp->m_dev,
+ agi_buf = libxfs_buf_get(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
mp->m_sb.sb_sectsize/BBSIZE);
agi_buf->b_ops = &xfs_agi_buf_ops;
@@ -1174,7 +1174,7 @@ build_ino_tree(xfs_mount_t *mp, xfs_agnumber_t agno,
lptr = &btree_curs->level[i];
agbno = get_next_blockaddr(agno, i, btree_curs);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
@@ -1306,7 +1306,7 @@ build_ino_tree(xfs_mount_t *mp, xfs_agnumber_t agno,
lptr->agbno = get_next_blockaddr(agno, 0, btree_curs);
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(lptr->agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, lptr->agbno),
XFS_FSB_TO_BB(mp, 1));
}
@@ -1459,7 +1459,7 @@ prop_rmap_cursor(
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
lptr->agbno = agbno;
@@ -1569,7 +1569,7 @@ build_rmap_tree(
lptr = &btree_curs->level[i];
agbno = get_next_blockaddr(agno, i, btree_curs);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
@@ -1668,7 +1668,7 @@ _("Insufficient memory to construct reverse-map cursor."));
lptr->agbno = get_next_blockaddr(agno, 0, btree_curs);
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(lptr->agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, lptr->agbno),
XFS_FSB_TO_BB(mp, 1));
}
@@ -1809,7 +1809,7 @@ prop_refc_cursor(
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
lptr->agbno = agbno;
@@ -1874,7 +1874,7 @@ build_refcount_tree(
lptr = &btree_curs->level[i];
agbno = get_next_blockaddr(agno, i, btree_curs);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1));
@@ -1961,7 +1961,7 @@ _("Insufficient memory to construct refcount cursor."));
lptr->agbno = get_next_blockaddr(agno, 0, btree_curs);
bt_hdr->bb_u.s.bb_rightsib = cpu_to_be32(lptr->agbno);
- lptr->buf_p = libxfs_getbuf(mp->m_dev,
+ lptr->buf_p = libxfs_buf_get(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, lptr->agbno),
XFS_FSB_TO_BB(mp, 1));
}
@@ -1996,7 +1996,7 @@ build_agf_agfl(
__be32 *freelist;
int error;
- agf_buf = libxfs_getbuf(mp->m_dev,
+ agf_buf = libxfs_buf_get(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
mp->m_sb.sb_sectsize/BBSIZE);
agf_buf->b_ops = &xfs_agf_buf_ops;
@@ -2068,7 +2068,7 @@ build_agf_agfl(
platform_uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid);
/* initialise the AGFL, then fill it if there are blocks left over. */
- agfl_buf = libxfs_getbuf(mp->m_dev,
+ agfl_buf = libxfs_buf_get(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGFL_DADDR(mp)),
mp->m_sb.sb_sectsize/BBSIZE);
agfl_buf->b_ops = &xfs_agfl_buf_ops;
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 06/18] libxfs: replace libxfs_readbuf with libxfs_buf_read
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (4 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 05/18] libxfs: replace libxfs_getbuf with libxfs_buf_get Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:44 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does Darrick J. Wong
` (11 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Change all the libxfs_readbuf calls to libxfs_buf_read to match the
kernel interface. This enables us to hide libxfs_readbuf and simplify
the userspace buffer interface further.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
copy/xfs_copy.c | 4 ++-
db/init.c | 2 +-
db/io.c | 2 +-
libxfs/init.c | 8 +++----
libxfs/libxfs_api_defs.h | 1 +
libxfs/libxfs_io.h | 25 +++++++++++++++------
libxfs/rdwr.c | 56 +++++++++++++++++++++++-----------------------
mkfs/xfs_mkfs.c | 4 ++-
repair/attr_repair.c | 6 ++---
repair/dino_chunks.c | 4 ++-
repair/dinode.c | 6 ++---
repair/phase3.c | 2 +-
repair/prefetch.c | 2 +-
repair/rt.c | 4 ++-
repair/scan.c | 12 +++++-----
15 files changed, 75 insertions(+), 63 deletions(-)
diff --git a/copy/xfs_copy.c b/copy/xfs_copy.c
index 74dd2e23..bb3ecd43 100644
--- a/copy/xfs_copy.c
+++ b/copy/xfs_copy.c
@@ -710,7 +710,7 @@ main(int argc, char **argv)
/* We don't yet know the sector size, so read maximal size */
libxfs_buftarg_init(&mbuf, xargs.ddev, xargs.logdev, xargs.rtdev);
- sbp = libxfs_readbuf(mbuf.m_ddev_targp, XFS_SB_DADDR,
+ sbp = libxfs_buf_read(mbuf.m_ddev_targp, XFS_SB_DADDR,
1 << (XFS_MAX_SECTORSIZE_LOG - BBSHIFT), 0, NULL);
sb = &mbuf.m_sb;
libxfs_sb_from_disk(sb, XFS_BUF_TO_SBP(sbp));
@@ -718,7 +718,7 @@ main(int argc, char **argv)
/* Do it again, now with proper length and verifier */
libxfs_buf_relse(sbp);
libxfs_purgebuf(sbp);
- sbp = libxfs_readbuf(mbuf.m_ddev_targp, XFS_SB_DADDR,
+ sbp = libxfs_buf_read(mbuf.m_ddev_targp, XFS_SB_DADDR,
1 << (sb->sb_sectlog - BBSHIFT),
0, &xfs_sb_buf_ops);
libxfs_buf_relse(sbp);
diff --git a/db/init.c b/db/init.c
index be4a08e5..acab349c 100644
--- a/db/init.c
+++ b/db/init.c
@@ -112,7 +112,7 @@ init(
*/
memset(&xmount, 0, sizeof(struct xfs_mount));
libxfs_buftarg_init(&xmount, x.ddev, x.logdev, x.rtdev);
- bp = libxfs_readbuf(xmount.m_ddev_targp, XFS_SB_DADDR,
+ bp = libxfs_buf_read(xmount.m_ddev_targp, XFS_SB_DADDR,
1 << (XFS_MAX_SECTORSIZE_LOG - BBSHIFT), 0, NULL);
if (!bp || bp->b_error) {
diff --git a/db/io.c b/db/io.c
index a11b7bb1..163ccc89 100644
--- a/db/io.c
+++ b/db/io.c
@@ -545,7 +545,7 @@ set_cur(
bp = libxfs_readbuf_map(mp->m_ddev_targp, bbmap->b,
bbmap->nmaps, 0, ops);
} else {
- bp = libxfs_readbuf(mp->m_ddev_targp, blknum, len, 0, ops);
+ bp = libxfs_buf_read(mp->m_ddev_targp, blknum, len, 0, ops);
iocur_top->bbmap = NULL;
}
diff --git a/libxfs/init.c b/libxfs/init.c
index 2c290428..c66cb785 100644
--- a/libxfs/init.c
+++ b/libxfs/init.c
@@ -438,7 +438,7 @@ rtmount_init(
(unsigned long long) mp->m_sb.sb_rblocks);
return -1;
}
- bp = libxfs_readbuf(mp->m_rtdev,
+ bp = libxfs_buf_read(mp->m_rtdev,
d - XFS_FSB_TO_BB(mp, 1), XFS_FSB_TO_BB(mp, 1), 0, NULL);
if (bp == NULL) {
fprintf(stderr, _("%s: realtime size check failed\n"),
@@ -721,7 +721,7 @@ libxfs_mount(
if (!(flags & LIBXFS_MOUNT_DEBUGGER))
readflags |= LIBXFS_READBUF_FAIL_EXIT;
- bp = libxfs_readbuf(mp->m_dev,
+ bp = libxfs_buf_read(mp->m_dev,
d - XFS_FSS_TO_BB(mp, 1), XFS_FSS_TO_BB(mp, 1),
readflags, NULL);
if (!bp) {
@@ -735,7 +735,7 @@ libxfs_mount(
mp->m_logdev_targp->dev != mp->m_ddev_targp->dev) {
d = (xfs_daddr_t) XFS_FSB_TO_BB(mp, mp->m_sb.sb_logblocks);
if ( (XFS_BB_TO_FSB(mp, d) != mp->m_sb.sb_logblocks) ||
- (!(bp = libxfs_readbuf(mp->m_logdev_targp,
+ (!(bp = libxfs_buf_read(mp->m_logdev_targp,
d - XFS_FSB_TO_BB(mp, 1),
XFS_FSB_TO_BB(mp, 1),
readflags, NULL))) ) {
@@ -763,7 +763,7 @@ libxfs_mount(
* read the first one and let the user know to check the geometry.
*/
if (sbp->sb_agcount > 1000000) {
- bp = libxfs_readbuf(mp->m_dev,
+ bp = libxfs_buf_read(mp->m_dev,
XFS_AG_DADDR(mp, sbp->sb_agcount - 1, 0), 1,
readflags, NULL);
if (bp->b_error) {
diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 6d86774f..57cf5f83 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -45,6 +45,7 @@
#define xfs_btree_init_block libxfs_btree_init_block
#define xfs_buf_delwri_submit libxfs_buf_delwri_submit
#define xfs_buf_get libxfs_buf_get
+#define xfs_buf_read libxfs_buf_read
#define xfs_buf_relse libxfs_buf_relse
#define xfs_bunmapi libxfs_bunmapi
#define xfs_calc_dquots_per_chunk libxfs_calc_dquots_per_chunk
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 5087f03d..9484c627 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -132,7 +132,7 @@ extern struct cache_operations libxfs_bcache_operations;
#ifdef XFS_BUF_TRACING
-#define libxfs_readbuf(dev, daddr, len, flags, ops) \
+#define libxfs_buf_read(dev, daddr, len, flags, ops) \
libxfs_trace_readbuf(__FUNCTION__, __FILE__, __LINE__, \
(dev), (daddr), (len), (flags), (ops))
#define libxfs_readbuf_map(dev, map, nmaps, flags, ops) \
@@ -153,9 +153,9 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_buf_relse(buf) \
libxfs_trace_putbuf(__FUNCTION__, __FILE__, __LINE__, (buf))
-extern xfs_buf_t *libxfs_trace_readbuf(const char *, const char *, int,
- struct xfs_buftarg *, xfs_daddr_t, int, int,
- const struct xfs_buf_ops *);
+struct xfs_buf *libxfs_trace_readbuf(const char *func, const char *file,
+ int line, struct xfs_buftarg *btp, xfs_daddr_t daddr,
+ size_t len, int flags, const struct xfs_buf_ops *ops);
extern xfs_buf_t *libxfs_trace_readbuf_map(const char *, const char *, int,
struct xfs_buftarg *, struct xfs_buf_map *, int, int,
const struct xfs_buf_ops *);
@@ -173,8 +173,6 @@ extern void libxfs_trace_putbuf (const char *, const char *, int,
#else
-extern xfs_buf_t *libxfs_readbuf(struct xfs_buftarg *, xfs_daddr_t, int, int,
- const struct xfs_buf_ops *);
extern xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
extern int libxfs_writebuf(xfs_buf_t *, int);
@@ -195,11 +193,24 @@ libxfs_buf_get(
return libxfs_getbuf_map(target, &map, 1, 0);
}
+static inline struct xfs_buf*
+libxfs_buf_read(
+ struct xfs_buftarg *target,
+ xfs_daddr_t blkno,
+ size_t numblks,
+ xfs_buf_flags_t flags,
+ const struct xfs_buf_ops *ops)
+{
+ DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
+
+ return libxfs_readbuf_map(target, &map, 1, flags, ops);
+}
+
#endif /* XFS_BUF_TRACING */
extern void libxfs_readbuf_verify(struct xfs_buf *bp,
const struct xfs_buf_ops *ops);
-extern xfs_buf_t *libxfs_getsb(struct xfs_mount *);
+struct xfs_buf *libxfs_getsb(struct xfs_mount *);
extern void libxfs_bcache_purge(void);
extern void libxfs_bcache_free(void);
extern void libxfs_bcache_flush(void);
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 56c50d47..ea2f4a13 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -44,7 +44,7 @@
*
* IOWs, userspace is behaving quite differently to the kernel and as a result
* it leaks errors from reads, invalidations and writes through
- * libxfs_buf_get/libxfs_readbuf.
+ * libxfs_buf_get/libxfs_buf_read.
*
* The result of this is that until the userspace code outside libxfs is cleaned
* up, functions that release buffers from userspace control (i.e
@@ -381,14 +381,11 @@ libxfs_log_header(
#ifdef XFS_BUF_TRACING
-#undef libxfs_readbuf
#undef libxfs_readbuf_map
#undef libxfs_writebuf
#undef libxfs_getbuf_map
#undef libxfs_getbuf_flags
-xfs_buf_t *libxfs_readbuf(struct xfs_buftarg *, xfs_daddr_t, int, int,
- const struct xfs_buf_ops *);
xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
int libxfs_writebuf(xfs_buf_t *, int);
@@ -409,22 +406,21 @@ do { \
} \
} while (0)
-xfs_buf_t *
-libxfs_trace_readbuf(const char *func, const char *file, int line,
- struct xfs_buftarg *btp, xfs_daddr_t blkno, int len, int flags,
- const struct xfs_buf_ops *ops)
+struct xfs_buf *
+libxfs_trace_readbuf(
+ const char *func,
+ const char *file,
+ int line,
+ struct xfs_buftarg *btp,
+ xfs_daddr_t blkno,
+ size_t len,
+ int flags,
+ const struct xfs_buf_ops *ops)
{
- xfs_buf_t *bp = libxfs_readbuf(btp, blkno, len, flags, ops);
- __add_trace(bp, func, file, line);
- return bp;
-}
+ struct xfs_buf *bp;
+ DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
-xfs_buf_t *
-libxfs_trace_readbuf_map(const char *func, const char *file, int line,
- struct xfs_buftarg *btp, struct xfs_buf_map *map, int nmaps, int flags,
- const struct xfs_buf_ops *ops)
-{
- xfs_buf_t *bp = libxfs_readbuf_map(btp, map, nmaps, flags, ops);
+ bp = libxfs_readbuf_map(btp, &map, 1, flags, ops);
__add_trace(bp, func, file, line);
return bp;
}
@@ -483,11 +479,12 @@ libxfs_trace_putbuf(const char *func, const char *file, int line, xfs_buf_t *bp)
#endif
-xfs_buf_t *
-libxfs_getsb(xfs_mount_t *mp)
+struct xfs_buf *
+libxfs_getsb(
+ struct xfs_mount *mp)
{
- return libxfs_readbuf(mp->m_ddev_targp, XFS_SB_DADDR,
- XFS_FSS_TO_BB(mp, 1), 0, &xfs_sb_buf_ops);
+ return libxfs_buf_read(mp->m_ddev_targp, XFS_SB_DADDR,
+ XFS_FSS_TO_BB(mp, 1), 0, &xfs_sb_buf_ops);
}
kmem_zone_t *xfs_buf_zone;
@@ -956,13 +953,16 @@ libxfs_readbuf_verify(struct xfs_buf *bp, const struct xfs_buf_ops *ops)
bp->b_flags &= ~LIBXFS_B_UNCHECKED;
}
-
-xfs_buf_t *
-libxfs_readbuf(struct xfs_buftarg *btp, xfs_daddr_t blkno, int len, int flags,
- const struct xfs_buf_ops *ops)
+static struct xfs_buf *
+libxfs_readbuf(
+ struct xfs_buftarg *btp,
+ xfs_daddr_t blkno,
+ size_t len,
+ int flags,
+ const struct xfs_buf_ops *ops)
{
- xfs_buf_t *bp;
- int error;
+ struct xfs_buf *bp;
+ int error;
bp = libxfs_getbuf_flags(btp, blkno, len, 0);
if (!bp)
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 948749e9..9ca4cb1a 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -3573,7 +3573,7 @@ rewrite_secondary_superblocks(
struct xfs_buf *buf;
/* rewrite the last superblock */
- buf = libxfs_readbuf(mp->m_dev,
+ buf = libxfs_buf_read(mp->m_dev,
XFS_AGB_TO_DADDR(mp, mp->m_sb.sb_agcount - 1,
XFS_SB_DADDR),
XFS_FSS_TO_BB(mp, 1),
@@ -3585,7 +3585,7 @@ rewrite_secondary_superblocks(
if (mp->m_sb.sb_agcount <= 2)
return;
- buf = libxfs_readbuf(mp->m_dev,
+ buf = libxfs_buf_read(mp->m_dev,
XFS_AGB_TO_DADDR(mp, (mp->m_sb.sb_agcount - 1) / 2,
XFS_SB_DADDR),
XFS_FSS_TO_BB(mp, 1),
diff --git a/repair/attr_repair.c b/repair/attr_repair.c
index ec068ba2..cc20b1a1 100644
--- a/repair/attr_repair.c
+++ b/repair/attr_repair.c
@@ -405,7 +405,7 @@ rmtval_get(xfs_mount_t *mp, xfs_ino_t ino, blkmap_t *blkmap,
clearit = 1;
break;
}
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
XFS_FSB_TO_BB(mp, 1), 0,
&xfs_attr3_rmt_buf_ops);
if (!bp) {
@@ -763,7 +763,7 @@ process_leaf_attr_level(xfs_mount_t *mp,
goto error_out;
}
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, dev_bno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, dev_bno),
XFS_FSB_TO_BB(mp, 1), 0,
&xfs_attr3_leaf_buf_ops);
if (!bp) {
@@ -1093,7 +1093,7 @@ process_longform_attr(
return 1;
}
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
XFS_FSB_TO_BB(mp, 1), 0, &xfs_da3_node_buf_ops);
if (!bp) {
do_warn(
diff --git a/repair/dino_chunks.c b/repair/dino_chunks.c
index 863c4531..76e9f773 100644
--- a/repair/dino_chunks.c
+++ b/repair/dino_chunks.c
@@ -39,7 +39,7 @@ check_aginode_block(xfs_mount_t *mp,
* tree and we wouldn't be here and we stale the buffers out
* so no one else will overlap them.
*/
- bp = libxfs_readbuf(mp->m_dev, XFS_AGB_TO_DADDR(mp, agno, agbno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp, 1), 0, NULL);
if (!bp) {
do_warn(_("cannot read agbno (%u/%u), disk block %" PRId64 "\n"),
@@ -656,7 +656,7 @@ process_inode_chunk(
pftrace("about to read off %llu in AG %d",
XFS_AGB_TO_DADDR(mp, agno, agbno), agno);
- bplist[bp_index] = libxfs_readbuf(mp->m_dev,
+ bplist[bp_index] = libxfs_buf_read(mp->m_dev,
XFS_AGB_TO_DADDR(mp, agno, agbno),
XFS_FSB_TO_BB(mp,
M_IGEO(mp)->blocks_per_cluster),
diff --git a/repair/dinode.c b/repair/dinode.c
index 928698cb..848eac09 100644
--- a/repair/dinode.c
+++ b/repair/dinode.c
@@ -748,7 +748,7 @@ get_agino_buf(
cluster_agino, cluster_daddr, cluster_blks);
#endif
- bp = libxfs_readbuf(mp->m_dev, cluster_daddr, cluster_blks,
+ bp = libxfs_buf_read(mp->m_dev, cluster_daddr, cluster_blks,
0, &xfs_inode_buf_ops);
if (!bp) {
do_warn(_("cannot read inode (%u/%u), disk block %" PRIu64 "\n"),
@@ -1179,7 +1179,7 @@ process_quota_inode(
fsbno = blkmap_get(blkmap, qbno);
dqid = (xfs_dqid_t)qbno * dqperchunk;
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, fsbno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, fsbno),
dqchunklen, 0, &xfs_dquot_buf_ops);
if (!bp) {
do_warn(
@@ -1284,7 +1284,7 @@ _("cannot read inode %" PRIu64 ", file block %d, NULL disk block\n"),
byte_cnt = XFS_FSB_TO_B(mp, blk_cnt);
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, fsbno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, fsbno),
BTOBB(byte_cnt), 0, &xfs_symlink_buf_ops);
if (!bp) {
do_warn(
diff --git a/repair/phase3.c b/repair/phase3.c
index 1c6929ac..886acd1f 100644
--- a/repair/phase3.c
+++ b/repair/phase3.c
@@ -28,7 +28,7 @@ process_agi_unlinked(
xfs_agnumber_t i;
int agi_dirty = 0;
- bp = libxfs_readbuf(mp->m_dev,
+ bp = libxfs_buf_read(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
mp->m_sb.sb_sectsize/BBSIZE, 0, &xfs_agi_buf_ops);
if (!bp)
diff --git a/repair/prefetch.c b/repair/prefetch.c
index f7ea9c8f..12272932 100644
--- a/repair/prefetch.c
+++ b/repair/prefetch.c
@@ -271,7 +271,7 @@ pf_scan_lbtree(
xfs_buf_t *bp;
int rc;
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, dbno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, dbno),
XFS_FSB_TO_BB(mp, 1), 0, &xfs_bmbt_buf_ops);
if (!bp)
return 0;
diff --git a/repair/rt.c b/repair/rt.c
index 3319829c..b514998d 100644
--- a/repair/rt.c
+++ b/repair/rt.c
@@ -193,7 +193,7 @@ process_rtbitmap(xfs_mount_t *mp,
error = 1;
continue;
}
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
XFS_FSB_TO_BB(mp, 1), NULL);
if (!bp) {
do_warn(_("can't read block %d for rtbitmap inode\n"),
@@ -255,7 +255,7 @@ process_rtsummary(xfs_mount_t *mp,
error++;
continue;
}
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, bno),
XFS_FSB_TO_BB(mp, 1), NULL);
if (!bp) {
do_warn(_("can't read block %d for rtsummary inode\n"),
diff --git a/repair/scan.c b/repair/scan.c
index c961e843..f4e4fef5 100644
--- a/repair/scan.c
+++ b/repair/scan.c
@@ -68,7 +68,7 @@ scan_sbtree(
{
xfs_buf_t *bp;
- bp = libxfs_readbuf(mp->m_dev, XFS_AGB_TO_DADDR(mp, agno, root),
+ bp = libxfs_buf_read(mp->m_dev, XFS_AGB_TO_DADDR(mp, agno, root),
XFS_FSB_TO_BB(mp, 1), 0, ops);
if (!bp) {
do_error(_("can't read btree block %d/%d\n"), agno, root);
@@ -123,7 +123,7 @@ scan_lbtree(
int dirty = 0;
bool badcrc = false;
- bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, root),
+ bp = libxfs_buf_read(mp->m_dev, XFS_FSB_TO_DADDR(mp, root),
XFS_FSB_TO_BB(mp, 1), 0, ops);
if (!bp) {
do_error(_("can't read btree block %d/%d\n"),
@@ -2111,7 +2111,7 @@ scan_freelist(
if (be32_to_cpu(agf->agf_flcount) == 0)
return;
- agflbuf = libxfs_readbuf(mp->m_dev,
+ agflbuf = libxfs_buf_read(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGFL_DADDR(mp)),
XFS_FSS_TO_BB(mp, 1), 0, &xfs_agfl_buf_ops);
if (!agflbuf) {
@@ -2335,7 +2335,7 @@ scan_ag(
return;
}
- sbbuf = libxfs_readbuf(mp->m_dev, XFS_AG_DADDR(mp, agno, XFS_SB_DADDR),
+ sbbuf = libxfs_buf_read(mp->m_dev, XFS_AG_DADDR(mp, agno, XFS_SB_DADDR),
XFS_FSS_TO_BB(mp, 1), 0, &xfs_sb_buf_ops);
if (!sbbuf) {
objname = _("root superblock");
@@ -2343,7 +2343,7 @@ scan_ag(
}
libxfs_sb_from_disk(sb, XFS_BUF_TO_SBP(sbbuf));
- agfbuf = libxfs_readbuf(mp->m_dev,
+ agfbuf = libxfs_buf_read(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
XFS_FSS_TO_BB(mp, 1), 0, &xfs_agf_buf_ops);
if (!agfbuf) {
@@ -2352,7 +2352,7 @@ scan_ag(
}
agf = XFS_BUF_TO_AGF(agfbuf);
- agibuf = libxfs_readbuf(mp->m_dev,
+ agibuf = libxfs_buf_read(mp->m_dev,
XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
XFS_FSS_TO_BB(mp, 1), 0, &xfs_agi_buf_ops);
if (!agibuf) {
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (5 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 06/18] libxfs: replace libxfs_readbuf with libxfs_buf_read Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:45 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 08/18] libxfs: make libxfs_readbuf_verify return an error code Darrick J. Wong
` (10 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Make libxfs_bwrite equivalent to libxfs_writebufr. This makes it so
that libxfs bwrite calls write the buffer to disk immediately and
without double-freeing the buffer. However, we choose to consolidate
around the libxfs_bwrite name to match the kernel.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
db/io.c | 4 ++--
libxfs/libxfs_api_defs.h | 1 +
libxfs/libxfs_io.h | 2 +-
libxfs/libxfs_priv.h | 1 -
libxfs/rdwr.c | 11 ++++++-----
5 files changed, 10 insertions(+), 9 deletions(-)
diff --git a/db/io.c b/db/io.c
index 163ccc89..7c7a4624 100644
--- a/db/io.c
+++ b/db/io.c
@@ -426,7 +426,7 @@ write_cur_buf(void)
{
int ret;
- ret = -libxfs_writebufr(iocur_top->bp);
+ ret = -libxfs_bwrite(iocur_top->bp);
if (ret != 0)
dbprintf(_("write error: %s\n"), strerror(ret));
@@ -442,7 +442,7 @@ write_cur_bbs(void)
{
int ret;
- ret = -libxfs_writebufr(iocur_top->bp);
+ ret = -libxfs_bwrite(iocur_top->bp);
if (ret != 0)
dbprintf(_("write error: %s\n"), strerror(ret));
diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 57cf5f83..df267c98 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -48,6 +48,7 @@
#define xfs_buf_read libxfs_buf_read
#define xfs_buf_relse libxfs_buf_relse
#define xfs_bunmapi libxfs_bunmapi
+#define xfs_bwrite libxfs_bwrite
#define xfs_calc_dquots_per_chunk libxfs_calc_dquots_per_chunk
#define xfs_da3_node_hdr_from_disk libxfs_da3_node_hdr_from_disk
#define xfs_da_get_buf libxfs_da_get_buf
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 9484c627..f00ff8d3 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -222,7 +222,7 @@ extern xfs_buf_t *libxfs_getbufr(struct xfs_buftarg *, xfs_daddr_t, int);
extern void libxfs_putbufr(xfs_buf_t *);
extern int libxfs_writebuf_int(xfs_buf_t *, int);
-extern int libxfs_writebufr(struct xfs_buf *);
+int libxfs_bwrite(struct xfs_buf *);
extern int libxfs_readbufr(struct xfs_buftarg *, xfs_daddr_t, xfs_buf_t *, int, int);
extern int libxfs_readbufr_map(struct xfs_buftarg *, struct xfs_buf *, int);
diff --git a/libxfs/libxfs_priv.h b/libxfs/libxfs_priv.h
index 0f26120f..5d6dd063 100644
--- a/libxfs/libxfs_priv.h
+++ b/libxfs/libxfs_priv.h
@@ -377,7 +377,6 @@ roundup_64(uint64_t x, uint32_t y)
(len) = __bar; /* no set-but-unused warning */ \
NULL; \
})
-#define xfs_bwrite(bp) libxfs_writebuf((bp), 0)
#define xfs_buf_oneshot(bp) ((void) 0)
#define XBRW_READ LIBXFS_BREAD
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index ea2f4a13..2a4ef15a 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -1091,9 +1091,10 @@ __write_buf(int fd, void *buf, int len, off64_t offset, int flags)
}
int
-libxfs_writebufr(xfs_buf_t *bp)
+libxfs_bwrite(
+ struct xfs_buf *bp)
{
- int fd = libxfs_device_to_fd(bp->b_target->dev);
+ int fd = libxfs_device_to_fd(bp->b_target->dev);
/*
* we never write buffers that are marked stale. This indicates they
@@ -1303,7 +1304,7 @@ libxfs_bflush(
struct xfs_buf *bp = (struct xfs_buf *)node;
if (!bp->b_error && bp->b_flags & LIBXFS_B_DIRTY)
- return libxfs_writebufr(bp);
+ return libxfs_bwrite(bp);
return bp->b_error;
}
@@ -1311,7 +1312,7 @@ void
libxfs_putbufr(xfs_buf_t *bp)
{
if (bp->b_flags & LIBXFS_B_DIRTY)
- libxfs_writebufr(bp);
+ libxfs_bwrite(bp);
libxfs_brelse((struct cache_node *)bp);
}
@@ -1524,7 +1525,7 @@ xfs_buf_delwri_submit(
list_for_each_entry_safe(bp, n, buffer_list, b_list) {
list_del_init(&bp->b_list);
- error2 = libxfs_writebufr(bp);
+ error2 = libxfs_bwrite(bp);
if (!error)
error = error2;
libxfs_buf_relse(bp);
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 08/18] libxfs: make libxfs_readbuf_verify return an error code
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (6 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:46 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 09/18] libxfs: make libxfs_readbufr stash the error value in b_error Darrick J. Wong
` (9 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Return the bp->b_error from libxfs_readbuf_verify instead of making
callers check it.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 3 +--
libxfs/rdwr.c | 10 +++++++---
repair/prefetch.c | 5 +++--
3 files changed, 11 insertions(+), 7 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index f00ff8d3..b01f2896 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -208,8 +208,7 @@ libxfs_buf_read(
#endif /* XFS_BUF_TRACING */
-extern void libxfs_readbuf_verify(struct xfs_buf *bp,
- const struct xfs_buf_ops *ops);
+int libxfs_readbuf_verify(struct xfs_buf *bp, const struct xfs_buf_ops *ops);
struct xfs_buf *libxfs_getsb(struct xfs_mount *);
extern void libxfs_bcache_purge(void);
extern void libxfs_bcache_free(void);
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 2a4ef15a..24c5eaf6 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -943,14 +943,18 @@ libxfs_readbufr(struct xfs_buftarg *btp, xfs_daddr_t blkno, xfs_buf_t *bp,
return error;
}
-void
-libxfs_readbuf_verify(struct xfs_buf *bp, const struct xfs_buf_ops *ops)
+int
+libxfs_readbuf_verify(
+ struct xfs_buf *bp,
+ const struct xfs_buf_ops *ops)
{
if (!ops)
- return;
+ return bp->b_error;
+
bp->b_ops = ops;
bp->b_ops->verify_read(bp);
bp->b_flags &= ~LIBXFS_B_UNCHECKED;
+ return bp->b_error;
}
static struct xfs_buf *
diff --git a/repair/prefetch.c b/repair/prefetch.c
index 12272932..a3858f9a 100644
--- a/repair/prefetch.c
+++ b/repair/prefetch.c
@@ -400,9 +400,10 @@ pf_read_inode_dirs(
int icnt = 0;
int hasdir = 0;
int isadir;
+ int error;
- libxfs_readbuf_verify(bp, &xfs_inode_buf_ops);
- if (bp->b_error)
+ error = -libxfs_readbuf_verify(bp, &xfs_inode_buf_ops);
+ if (error)
return;
for (icnt = 0; icnt < (bp->b_bcount >> mp->m_sb.sb_inodelog); icnt++) {
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 09/18] libxfs: make libxfs_readbufr stash the error value in b_error
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (7 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 08/18] libxfs: make libxfs_readbuf_verify return an error code Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:46 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached Darrick J. Wong
` (8 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Make libxfs_readbufr stash the error value in b_error, which will make
the behavior consistent between regular and multi-mapping buffers.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/rdwr.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 24c5eaf6..51f93c3e 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -940,6 +940,7 @@ libxfs_readbufr(struct xfs_buftarg *btp, xfs_daddr_t blkno, xfs_buf_t *bp,
pthread_self(), __FUNCTION__, bytes, error,
(long long)LIBXFS_BBTOOFF64(blkno), (long long)blkno, bp);
#endif
+ bp->b_error = error;
return error;
}
@@ -999,9 +1000,7 @@ libxfs_readbuf(
* contents. *cough* xfs_da_node_buf_ops *cough*.
*/
error = libxfs_readbufr(btp, blkno, bp, len, flags);
- if (error)
- bp->b_error = error;
- else
+ if (!error)
libxfs_readbuf_verify(bp, ops);
return bp;
}
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (8 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 09/18] libxfs: make libxfs_readbufr stash the error value in b_error Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:48 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them Darrick J. Wong
` (7 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Introduce an uncached read function so that userspace can handle them in
the same way as the kernel. This also eliminates the need for some of
the libxfs_purgebuf calls (and two trips into the cache code).
Refactor the get/read uncached buffer functions to hide the details of
uncached buffer-ism in rdwr.c.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
copy/xfs_copy.c | 13 +++++++--
db/init.c | 9 +++---
libxfs/libxfs_api_defs.h | 2 +
libxfs/libxfs_io.h | 22 +++------------
libxfs/rdwr.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++
5 files changed, 88 insertions(+), 25 deletions(-)
diff --git a/copy/xfs_copy.c b/copy/xfs_copy.c
index bb3ecd43..75c39407 100644
--- a/copy/xfs_copy.c
+++ b/copy/xfs_copy.c
@@ -562,6 +562,7 @@ main(int argc, char **argv)
libxfs_init_t xargs;
thread_args *tcarg;
struct stat statbuf;
+ int error;
progname = basename(argv[0]);
@@ -710,14 +711,20 @@ main(int argc, char **argv)
/* We don't yet know the sector size, so read maximal size */
libxfs_buftarg_init(&mbuf, xargs.ddev, xargs.logdev, xargs.rtdev);
- sbp = libxfs_buf_read(mbuf.m_ddev_targp, XFS_SB_DADDR,
- 1 << (XFS_MAX_SECTORSIZE_LOG - BBSHIFT), 0, NULL);
+ error = -libxfs_buf_read_uncached(mbuf.m_ddev_targp, XFS_SB_DADDR,
+ 1 << (XFS_MAX_SECTORSIZE_LOG - BBSHIFT), 0, &sbp, NULL);
+ if (error) {
+ do_log(_("%s: couldn't read superblock, error=%d\n"),
+ progname, error);
+ exit(1);
+ }
+
sb = &mbuf.m_sb;
libxfs_sb_from_disk(sb, XFS_BUF_TO_SBP(sbp));
/* Do it again, now with proper length and verifier */
libxfs_buf_relse(sbp);
- libxfs_purgebuf(sbp);
+
sbp = libxfs_buf_read(mbuf.m_ddev_targp, XFS_SB_DADDR,
1 << (sb->sb_sectlog - BBSHIFT),
0, &xfs_sb_buf_ops);
diff --git a/db/init.c b/db/init.c
index acab349c..eb53d672 100644
--- a/db/init.c
+++ b/db/init.c
@@ -47,6 +47,7 @@ init(
struct xfs_buf *bp;
unsigned int agcount;
int c;
+ int error;
setlocale(LC_ALL, "");
bindtextdomain(PACKAGE, LOCALEDIR);
@@ -112,10 +113,9 @@ init(
*/
memset(&xmount, 0, sizeof(struct xfs_mount));
libxfs_buftarg_init(&xmount, x.ddev, x.logdev, x.rtdev);
- bp = libxfs_buf_read(xmount.m_ddev_targp, XFS_SB_DADDR,
- 1 << (XFS_MAX_SECTORSIZE_LOG - BBSHIFT), 0, NULL);
-
- if (!bp || bp->b_error) {
+ error = -libxfs_buf_read_uncached(xmount.m_ddev_targp, XFS_SB_DADDR,
+ 1 << (XFS_MAX_SECTORSIZE_LOG - BBSHIFT), 0, &bp, NULL);
+ if (error) {
fprintf(stderr, _("%s: %s is invalid (cannot read first 512 "
"bytes)\n"), progname, fsdevice);
exit(1);
@@ -124,7 +124,6 @@ init(
/* copy SB from buffer to in-core, converting architecture as we go */
libxfs_sb_from_disk(&xmount.m_sb, XFS_BUF_TO_SBP(bp));
libxfs_buf_relse(bp);
- libxfs_purgebuf(bp);
sbp = &xmount.m_sb;
if (sbp->sb_magicnum != XFS_SB_MAGIC) {
diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index df267c98..1149e301 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -45,7 +45,9 @@
#define xfs_btree_init_block libxfs_btree_init_block
#define xfs_buf_delwri_submit libxfs_buf_delwri_submit
#define xfs_buf_get libxfs_buf_get
+#define xfs_buf_get_uncached libxfs_buf_get_uncached
#define xfs_buf_read libxfs_buf_read
+#define xfs_buf_read_uncached libxfs_buf_read_uncached
#define xfs_buf_relse libxfs_buf_relse
#define xfs_bunmapi libxfs_bunmapi
#define xfs_bwrite libxfs_bwrite
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index b01f2896..546b7710 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -257,23 +257,11 @@ xfs_buf_associate_memory(struct xfs_buf *bp, void *mem, size_t len)
return 0;
}
-/*
- * Allocate an uncached buffer that points nowhere. The refcount will be 1,
- * and the cache node hash list will be empty to indicate that it's uncached.
- */
-static inline struct xfs_buf *
-xfs_buf_get_uncached(struct xfs_buftarg *targ, size_t bblen, int flags)
-{
- struct xfs_buf *bp;
-
- bp = libxfs_getbufr(targ, XFS_BUF_DADDR_NULL, bblen);
- if (!bp)
- return NULL;
-
- INIT_LIST_HEAD(&bp->b_node.cn_hash);
- bp->b_node.cn_count = 1;
- return bp;
-}
+struct xfs_buf *libxfs_buf_get_uncached(struct xfs_buftarg *targ, size_t bblen,
+ int flags);
+int libxfs_buf_read_uncached(struct xfs_buftarg *targ, xfs_daddr_t daddr,
+ size_t bblen, int flags, struct xfs_buf **bpp,
+ const struct xfs_buf_ops *ops);
/* Push a single buffer on a delwri queue. */
static inline void
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 51f93c3e..ada20dd9 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -1070,6 +1070,73 @@ libxfs_readbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map, int nmaps,
return bp;
}
+/* Allocate a raw uncached buffer. */
+static inline struct xfs_buf *
+libxfs_getbufr_uncached(
+ struct xfs_buftarg *targ,
+ xfs_daddr_t daddr,
+ size_t bblen)
+{
+ struct xfs_buf *bp;
+
+ bp = libxfs_getbufr(targ, daddr, bblen);
+ if (!bp)
+ return NULL;
+
+ INIT_LIST_HEAD(&bp->b_node.cn_hash);
+ bp->b_node.cn_count = 1;
+ return bp;
+}
+
+/*
+ * Allocate an uncached buffer that points nowhere. The refcount will be 1,
+ * and the cache node hash list will be empty to indicate that it's uncached.
+ */
+struct xfs_buf *
+libxfs_buf_get_uncached(
+ struct xfs_buftarg *targ,
+ size_t bblen,
+ int flags)
+{
+ return libxfs_getbufr_uncached(targ, XFS_BUF_DADDR_NULL, bblen);
+}
+
+/*
+ * Allocate and read an uncached buffer. The refcount will be 1, and the cache
+ * node hash list will be empty to indicate that it's uncached.
+ */
+int
+libxfs_buf_read_uncached(
+ struct xfs_buftarg *targ,
+ xfs_daddr_t daddr,
+ size_t bblen,
+ int flags,
+ struct xfs_buf **bpp,
+ const struct xfs_buf_ops *ops)
+{
+ struct xfs_buf *bp;
+ int error;
+
+ *bpp = NULL;
+ bp = libxfs_getbufr_uncached(targ, daddr, bblen);
+ if (!bp)
+ return -ENOMEM;
+
+ error = libxfs_readbufr(targ, daddr, bp, bblen, flags);
+ if (error)
+ goto err;
+
+ error = libxfs_readbuf_verify(bp, ops);
+ if (error)
+ goto err;
+
+ *bpp = bp;
+ return 0;
+err:
+ libxfs_buf_relse(bp);
+ return error;
+}
+
static int
__write_buf(int fd, void *buf, int len, off64_t offset, int flags)
{
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (9 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:49 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes Darrick J. Wong
` (6 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Use the new uncached buffer functions to manage buffers instead of
open-coding the logic.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
include/libxlog.h | 1 -
libxlog/xfs_log_recover.c | 35 ++++++++++++++---------------------
logprint/log_print_all.c | 2 +-
3 files changed, 15 insertions(+), 23 deletions(-)
diff --git a/include/libxlog.h b/include/libxlog.h
index 4b785507..5e94fa1e 100644
--- a/include/libxlog.h
+++ b/include/libxlog.h
@@ -75,7 +75,6 @@ extern libxfs_init_t x;
extern int xlog_is_dirty(struct xfs_mount *, struct xlog *, libxfs_init_t *,
int);
extern struct xfs_buf *xlog_get_bp(struct xlog *, int);
-extern void xlog_put_bp(struct xfs_buf *);
extern int xlog_bread(struct xlog *log, xfs_daddr_t blk_no, int nbblks,
xfs_buf_t *bp, char **offset);
extern int xlog_bread_noalign(struct xlog *log, xfs_daddr_t blk_no,
diff --git a/libxlog/xfs_log_recover.c b/libxlog/xfs_log_recover.c
index 9324a213..e7e57bd2 100644
--- a/libxlog/xfs_log_recover.c
+++ b/libxlog/xfs_log_recover.c
@@ -67,14 +67,7 @@ xlog_get_bp(
nbblks += log->l_sectBBsize;
nbblks = round_up(nbblks, log->l_sectBBsize);
- return libxfs_getbufr(log->l_dev, (xfs_daddr_t)-1, nbblks);
-}
-
-void
-xlog_put_bp(
- xfs_buf_t *bp)
-{
- libxfs_putbufr(bp);
+ return libxfs_buf_get_uncached(log->l_dev, nbblks, 0);
}
/*
@@ -274,7 +267,7 @@ xlog_find_verify_cycle(
*new_blk = -1;
out:
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
return error;
}
@@ -383,7 +376,7 @@ xlog_find_verify_log_record(
*last_blk = i;
out:
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
return error;
}
@@ -634,7 +627,7 @@ xlog_find_head(
goto bp_err;
}
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
if (head_blk == log_bbnum)
*return_head_blk = 0;
else
@@ -648,7 +641,7 @@ xlog_find_head(
return 0;
bp_err:
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
if (error)
xfs_warn(log->l_mp, "failed to find log head");
@@ -745,7 +738,7 @@ xlog_find_tail(
}
if (!found) {
xfs_warn(log->l_mp, "%s: couldn't find sync record", __func__);
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
ASSERT(0);
return XFS_ERROR(EIO);
}
@@ -858,7 +851,7 @@ xlog_find_tail(
error = xlog_clear_stale_blocks(log, tail_lsn);
done:
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
if (error)
xfs_warn(log->l_mp, "failed to locate log tail");
@@ -906,7 +899,7 @@ xlog_find_zeroed(
first_cycle = xlog_get_cycle(offset);
if (first_cycle == 0) { /* completely zeroed log */
*blk_no = 0;
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
return -1;
}
@@ -917,7 +910,7 @@ xlog_find_zeroed(
last_cycle = xlog_get_cycle(offset);
if (last_cycle != 0) { /* log completely written to */
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
return 0;
} else if (first_cycle != 1) {
/*
@@ -974,7 +967,7 @@ xlog_find_zeroed(
*blk_no = last_blk;
bp_err:
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
if (error)
return error;
return -1;
@@ -1457,7 +1450,7 @@ xlog_do_recovery_pass(
hblks = h_size / XLOG_HEADER_CYCLE_SIZE;
if (h_size % XLOG_HEADER_CYCLE_SIZE)
hblks++;
- xlog_put_bp(hbp);
+ libxfs_buf_relse(hbp);
hbp = xlog_get_bp(log, hblks);
} else {
hblks = 1;
@@ -1473,7 +1466,7 @@ xlog_do_recovery_pass(
return ENOMEM;
dbp = xlog_get_bp(log, BTOBB(h_size));
if (!dbp) {
- xlog_put_bp(hbp);
+ libxfs_buf_relse(hbp);
return ENOMEM;
}
@@ -1657,8 +1650,8 @@ xlog_do_recovery_pass(
}
bread_err2:
- xlog_put_bp(dbp);
+ libxfs_buf_relse(dbp);
bread_err1:
- xlog_put_bp(hbp);
+ libxfs_buf_relse(hbp);
return error;
}
diff --git a/logprint/log_print_all.c b/logprint/log_print_all.c
index d3d4c07b..32d13719 100644
--- a/logprint/log_print_all.c
+++ b/logprint/log_print_all.c
@@ -39,7 +39,7 @@ xlog_print_find_oldest(
error = xlog_find_cycle_start(log, bp, first_blk,
last_blk, last_half_cycle);
- xlog_put_bp(bp);
+ libxfs_buf_relse(bp);
return error;
}
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (10 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:51 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion Darrick J. Wong
` (5 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Teach mkfs to use uncached buffers to write the start and end of the
data device, the initial superblock, and the end of the realtime device
instead of open-coding uncached buffers. This means we can get rid of
libxfs_purgebuf since we handle the state from the start now.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 1 -
libxfs/rdwr.c | 12 ------------
mkfs/xfs_mkfs.c | 34 +++++++++++++++++++++++-----------
3 files changed, 23 insertions(+), 24 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 546b7710..6598dba7 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -213,7 +213,6 @@ struct xfs_buf *libxfs_getsb(struct xfs_mount *);
extern void libxfs_bcache_purge(void);
extern void libxfs_bcache_free(void);
extern void libxfs_bcache_flush(void);
-extern void libxfs_purgebuf(xfs_buf_t *);
extern int libxfs_bcache_overflowed(void);
/* Buffer (Raw) Interfaces */
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index ada20dd9..20a8b0ce 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -869,18 +869,6 @@ libxfs_buf_relse(
libxfs_putbufr(bp);
}
-void
-libxfs_purgebuf(xfs_buf_t *bp)
-{
- struct xfs_bufkey key = {NULL};
-
- key.buftarg = bp->b_target;
- key.blkno = bp->b_bn;
- key.bblen = bp->b_length;
-
- cache_node_purge(libxfs_bcache, &key, (struct cache_node *)bp);
-}
-
static struct cache_node *
libxfs_balloc(cache_key_t key)
{
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 9ca4cb1a..f58f235d 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -3396,6 +3396,21 @@ finish_superblock_setup(
}
+/* Prepare an uncached buffer, ready to write something out. */
+static inline struct xfs_buf *
+get_write_buf(
+ struct xfs_buftarg *btp,
+ xfs_daddr_t daddr,
+ int bblen)
+{
+ struct xfs_buf *bp;
+
+ bp = libxfs_buf_get_uncached(btp, bblen, 0);
+ bp->b_bn = daddr;
+ bp->b_maps[0].bm_bn = daddr;
+ return bp;
+}
+
/*
* Sanitise the data and log devices and prepare them so libxfs can mount the
* device successfully. Also check we can access the rt device if configured.
@@ -3444,11 +3459,10 @@ prepare_devices(
* the end of the device. (MD sb is ~64k from the end, take out a wider
* swath to be sure)
*/
- buf = libxfs_buf_get(mp->m_ddev_targp, (xi->dsize - whack_blks),
- whack_blks);
+ buf = get_write_buf(mp->m_ddev_targp, (xi->dsize - whack_blks),
+ whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
- libxfs_purgebuf(buf);
/*
* Now zero out the beginning of the device, to obliterate any old
@@ -3456,18 +3470,17 @@ prepare_devices(
* swap (somewhere around the page size), jfs (32k),
* ext[2,3] and reiserfs (64k) - and hopefully all else.
*/
- buf = libxfs_buf_get(mp->m_ddev_targp, 0, whack_blks);
+ buf = get_write_buf(mp->m_ddev_targp, 0, whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
- libxfs_purgebuf(buf);
/* OK, now write the superblock... */
- buf = libxfs_buf_get(mp->m_ddev_targp, XFS_SB_DADDR, XFS_FSS_TO_BB(mp, 1));
+ buf = get_write_buf(mp->m_ddev_targp, XFS_SB_DADDR,
+ XFS_FSS_TO_BB(mp, 1));
buf->b_ops = &xfs_sb_buf_ops;
memset(buf->b_addr, 0, cfg->sectorsize);
libxfs_sb_to_disk(buf->b_addr, sbp);
libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
- libxfs_purgebuf(buf);
/* ...and zero the log.... */
lsunit = sbp->sb_logsunit;
@@ -3482,12 +3495,11 @@ prepare_devices(
/* finally, check we can write the last block in the realtime area */
if (mp->m_rtdev_targp->dev && cfg->rtblocks > 0) {
- buf = libxfs_buf_get(mp->m_rtdev_targp,
- XFS_FSB_TO_BB(mp, cfg->rtblocks - 1LL),
- BTOBB(cfg->blocksize));
+ buf = get_write_buf(mp->m_rtdev_targp,
+ XFS_FSB_TO_BB(mp, cfg->rtblocks - 1LL),
+ BTOBB(cfg->blocksize));
memset(buf->b_addr, 0, cfg->blocksize);
libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
- libxfs_purgebuf(buf);
}
}
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (11 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes Darrick J. Wong
@ 2020-02-20 1:43 ` Darrick J. Wong
2020-02-21 14:53 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 14/18] libxfs: remove libxfs_writebuf_int Darrick J. Wong
` (4 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:43 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
libxfs_writebuf is not a well named function -- it marks the buffer
dirty and then releases the caller's reference. The actual write comes
when the cache is flushed, either because someone explicitly told the
cache to flush or because we started buffer reclaim.
Make the buffer release explicit in the callers and rename the function
to say what it actually does -- it marks the buffer dirty outside of
transaction context.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 10 +++++-----
libxfs/rdwr.c | 21 ++++++++++++++-------
libxfs/trans.c | 3 ++-
mkfs/proto.c | 6 ++++--
mkfs/xfs_mkfs.c | 21 ++++++++++++++-------
repair/attr_repair.c | 17 +++++++++--------
repair/da_util.c | 12 ++++++++----
repair/dino_chunks.c | 6 ++++--
repair/dinode.c | 12 ++++++++----
repair/dir2.c | 9 ++++++---
repair/phase3.c | 6 ++++--
repair/phase5.c | 42 ++++++++++++++++++++++++++++--------------
repair/rmap.c | 3 ++-
repair/scan.c | 21 ++++++++++++++-------
repair/xfs_repair.c | 3 ++-
15 files changed, 124 insertions(+), 68 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 6598dba7..ad8acc1e 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -138,8 +138,8 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_readbuf_map(dev, map, nmaps, flags, ops) \
libxfs_trace_readbuf_map(__FUNCTION__, __FILE__, __LINE__, \
(dev), (map), (nmaps), (flags), (ops))
-#define libxfs_writebuf(buf, flags) \
- libxfs_trace_writebuf(__FUNCTION__, __FILE__, __LINE__, \
+#define libxfs_buf_dirty(buf, flags) \
+ libxfs_trace_dirtybuf(__FUNCTION__, __FILE__, __LINE__, \
(buf), (flags))
#define libxfs_buf_get(dev, daddr, len) \
libxfs_trace_getbuf(__FUNCTION__, __FILE__, __LINE__, \
@@ -159,8 +159,8 @@ struct xfs_buf *libxfs_trace_readbuf(const char *func, const char *file,
extern xfs_buf_t *libxfs_trace_readbuf_map(const char *, const char *, int,
struct xfs_buftarg *, struct xfs_buf_map *, int, int,
const struct xfs_buf_ops *);
-extern int libxfs_trace_writebuf(const char *, const char *, int,
- xfs_buf_t *, int);
+void libxfs_trace_dirtybuf(const char *func, const char *file, int line,
+ struct xfs_buf *bp, int flags);
struct xfs_buf *libxfs_trace_getbuf(const char *func, const char *file,
int line, struct xfs_buftarg *btp, xfs_daddr_t daddr,
size_t len);
@@ -175,7 +175,7 @@ extern void libxfs_trace_putbuf (const char *, const char *, int,
extern xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
-extern int libxfs_writebuf(xfs_buf_t *, int);
+void libxfs_buf_dirty(struct xfs_buf *bp, int flags);
extern xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *,
struct xfs_buf_map *, int, int);
extern xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t,
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 20a8b0ce..af363bef 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -425,11 +425,16 @@ libxfs_trace_readbuf(
return bp;
}
-int
-libxfs_trace_writebuf(const char *func, const char *file, int line, xfs_buf_t *bp, int flags)
+void
+libxfs_trace_dirtybuf(
+ const char *func,
+ const char *file,
+ int line,
+ struct xfs_buf *bp,
+ int flags)
{
__add_trace(bp, func, file, line);
- return libxfs_writebuf(bp, flags);
+ libxfs_buf_dirty(bp, flags);
}
struct xfs_buf *
@@ -1233,8 +1238,12 @@ libxfs_writebuf_int(xfs_buf_t *bp, int flags)
return 0;
}
-int
-libxfs_writebuf(
+/*
+ * Mark a buffer dirty. The dirty data will be written out when the cache
+ * is flushed (or at release time if the buffer is uncached).
+ */
+void
+libxfs_buf_dirty(
struct xfs_buf *bp,
int flags)
{
@@ -1256,8 +1265,6 @@ libxfs_writebuf(
bp->b_error = 0;
bp->b_flags &= ~LIBXFS_B_STALE;
bp->b_flags |= bflags;
- libxfs_buf_relse(bp);
- return 0;
}
void
diff --git a/libxfs/trans.c b/libxfs/trans.c
index 59cb897f..4c208422 100644
--- a/libxfs/trans.c
+++ b/libxfs/trans.c
@@ -843,7 +843,8 @@ inode_item_done(
goto free;
}
- libxfs_writebuf(bp, 0);
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
free:
xfs_inode_item_put(iip);
}
diff --git a/mkfs/proto.c b/mkfs/proto.c
index 0025fa08..de5ae306 100644
--- a/mkfs/proto.c
+++ b/mkfs/proto.c
@@ -261,8 +261,10 @@ newfile(
memset((char *)bp->b_addr + len, 0, bp->b_bcount - len);
if (logit)
libxfs_trans_log_buf(tp, bp, 0, bp->b_bcount - 1);
- else
- libxfs_writebuf(bp, LIBXFS_WRITEBUF_FAIL_EXIT);
+ else {
+ libxfs_buf_dirty(bp, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(bp);
+ }
}
ip->i_d.di_size = len;
return flags;
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index f58f235d..b50b8b3a 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -3462,7 +3462,8 @@ prepare_devices(
buf = get_write_buf(mp->m_ddev_targp, (xi->dsize - whack_blks),
whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
/*
* Now zero out the beginning of the device, to obliterate any old
@@ -3472,7 +3473,8 @@ prepare_devices(
*/
buf = get_write_buf(mp->m_ddev_targp, 0, whack_blks);
memset(buf->b_addr, 0, WHACK_SIZE);
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
/* OK, now write the superblock... */
buf = get_write_buf(mp->m_ddev_targp, XFS_SB_DADDR,
@@ -3480,7 +3482,8 @@ prepare_devices(
buf->b_ops = &xfs_sb_buf_ops;
memset(buf->b_addr, 0, cfg->sectorsize);
libxfs_sb_to_disk(buf->b_addr, sbp);
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
/* ...and zero the log.... */
lsunit = sbp->sb_logsunit;
@@ -3499,7 +3502,8 @@ prepare_devices(
XFS_FSB_TO_BB(mp, cfg->rtblocks - 1LL),
BTOBB(cfg->blocksize));
memset(buf->b_addr, 0, cfg->blocksize);
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
}
}
@@ -3591,7 +3595,8 @@ rewrite_secondary_superblocks(
XFS_FSS_TO_BB(mp, 1),
LIBXFS_READBUF_FAIL_EXIT, &xfs_sb_buf_ops);
XFS_BUF_TO_SBP(buf)->sb_rootino = cpu_to_be64(mp->m_sb.sb_rootino);
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
/* and one in the middle for luck if there's enough AGs for that */
if (mp->m_sb.sb_agcount <= 2)
@@ -3603,7 +3608,8 @@ rewrite_secondary_superblocks(
XFS_FSS_TO_BB(mp, 1),
LIBXFS_READBUF_FAIL_EXIT, &xfs_sb_buf_ops);
XFS_BUF_TO_SBP(buf)->sb_rootino = cpu_to_be64(mp->m_sb.sb_rootino);
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
}
static void
@@ -3951,7 +3957,8 @@ main(
if (!buf || buf->b_error)
exit(1);
(XFS_BUF_TO_SBP(buf))->sb_inprogress = 0;
- libxfs_writebuf(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_dirty(buf, LIBXFS_WRITEBUF_FAIL_EXIT);
+ libxfs_buf_relse(buf);
/* Make sure our new fs made it to stable storage. */
libxfs_flush_devices(mp, &d, &l, &r);
diff --git a/repair/attr_repair.c b/repair/attr_repair.c
index cc20b1a1..fb64b0be 100644
--- a/repair/attr_repair.c
+++ b/repair/attr_repair.c
@@ -835,8 +835,10 @@ process_leaf_attr_level(xfs_mount_t *mp,
if (!no_modify && bp->b_error == -EFSBADCRC)
repair++;
- if (repair && !no_modify)
- libxfs_writebuf(bp, 0);
+ if (repair && !no_modify) {
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
+ }
else
libxfs_buf_relse(bp);
} while (da_bno != 0);
@@ -999,9 +1001,8 @@ _("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "
*repair = *repair || repairlinks;
if (*repair && !no_modify)
- libxfs_writebuf(bp, 0);
- else
- libxfs_buf_relse(bp);
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
return 0;
}
@@ -1043,9 +1044,9 @@ _("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "
/* must do this now, to release block 0 before the traversal */
if ((*repair || repairlinks) && !no_modify) {
*repair = 1;
- libxfs_writebuf(bp, 0);
- } else
- libxfs_buf_relse(bp);
+ libxfs_buf_dirty(bp, 0);
+ }
+ libxfs_buf_relse(bp);
error = process_node_attr(mp, ino, dip, blkmap); /* + repair */
if (error)
*repair = 0;
diff --git a/repair/da_util.c b/repair/da_util.c
index c02d621c..d1e17ec3 100644
--- a/repair/da_util.c
+++ b/repair/da_util.c
@@ -403,8 +403,10 @@ _("would correct bad hashval in non-leaf %s block\n"
ASSERT(cursor->level[this_level].dirty == 0 ||
(cursor->level[this_level].dirty && !no_modify));
- if (cursor->level[this_level].dirty && !no_modify)
- libxfs_writebuf(cursor->level[this_level].bp, 0);
+ if (cursor->level[this_level].dirty && !no_modify) {
+ libxfs_buf_dirty(cursor->level[this_level].bp, 0);
+ libxfs_buf_relse(cursor->level[this_level].bp);
+ }
else
libxfs_buf_relse(cursor->level[this_level].bp);
@@ -619,8 +621,10 @@ _("bad level %d in %s block %u for inode %" PRIu64 "\n"),
cursor->level[this_level].bp->b_error == -EFSBADCRC)
cursor->level[this_level].dirty = 1;
- if (cursor->level[this_level].dirty && !no_modify)
- libxfs_writebuf(cursor->level[this_level].bp, 0);
+ if (cursor->level[this_level].dirty && !no_modify) {
+ libxfs_buf_dirty(cursor->level[this_level].bp, 0);
+ libxfs_buf_relse(cursor->level[this_level].bp);
+ }
else
libxfs_buf_relse(cursor->level[this_level].bp);
diff --git a/repair/dino_chunks.c b/repair/dino_chunks.c
index 76e9f773..1378fc59 100644
--- a/repair/dino_chunks.c
+++ b/repair/dino_chunks.c
@@ -939,8 +939,10 @@ process_inode_chunk(
bplist[bp_index], (long long)
XFS_BUF_ADDR(bplist[bp_index]), agno);
- if (dirty && !no_modify)
- libxfs_writebuf(bplist[bp_index], 0);
+ if (dirty && !no_modify) {
+ libxfs_buf_dirty(bplist[bp_index], 0);
+ libxfs_buf_relse(bplist[bp_index]);
+ }
else
libxfs_buf_relse(bplist[bp_index]);
}
diff --git a/repair/dinode.c b/repair/dinode.c
index 848eac09..276dd2e1 100644
--- a/repair/dinode.c
+++ b/repair/dinode.c
@@ -1231,8 +1231,10 @@ _("cannot read inode %" PRIu64 ", file block %" PRIu64 ", disk block %" PRIu64 "
}
}
- if (writebuf && !no_modify)
- libxfs_writebuf(bp, 0);
+ if (writebuf && !no_modify) {
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
+ }
else
libxfs_buf_relse(bp);
}
@@ -1328,8 +1330,10 @@ _("bad symlink header ino %" PRIu64 ", file block %d, disk block %" PRIu64 "\n")
offset += byte_cnt;
i++;
- if (badcrc && !no_modify)
- libxfs_writebuf(bp, 0);
+ if (badcrc && !no_modify) {
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
+ }
else
libxfs_buf_relse(bp);
}
diff --git a/repair/dir2.c b/repair/dir2.c
index 769e341c..a8b8de58 100644
--- a/repair/dir2.c
+++ b/repair/dir2.c
@@ -1010,7 +1010,8 @@ _("bad directory block magic # %#x in block %u for directory inode %" PRIu64 "\n
dirty = 1;
if (dirty && !no_modify) {
*repair = 1;
- libxfs_writebuf(bp, 0);
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
} else
libxfs_buf_relse(bp);
return rval;
@@ -1180,7 +1181,8 @@ _("bad sibling back pointer for block %u in directory inode %" PRIu64 "\n"),
ASSERT(buf_dirty == 0 || (buf_dirty && !no_modify));
if (buf_dirty && !no_modify) {
*repair = 1;
- libxfs_writebuf(bp, 0);
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
} else
libxfs_buf_relse(bp);
} while (da_bno != 0);
@@ -1339,7 +1341,8 @@ _("bad directory block magic # %#x in block %" PRIu64 " for directory inode %" P
}
if (dirty && !no_modify) {
*repair = 1;
- libxfs_writebuf(bp, 0);
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
} else
libxfs_buf_relse(bp);
}
diff --git a/repair/phase3.c b/repair/phase3.c
index 886acd1f..396743a4 100644
--- a/repair/phase3.c
+++ b/repair/phase3.c
@@ -46,8 +46,10 @@ process_agi_unlinked(
}
}
- if (agi_dirty)
- libxfs_writebuf(bp, 0);
+ if (agi_dirty) {
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
+ }
else
libxfs_buf_relse(bp);
}
diff --git a/repair/phase5.c b/repair/phase5.c
index cdbf6697..561a6b3f 100644
--- a/repair/phase5.c
+++ b/repair/phase5.c
@@ -321,9 +321,11 @@ write_cursor(bt_status_t *curs)
fprintf(stderr, "writing bt prev block %u\n",
curs->level[i].prev_agbno);
#endif
- libxfs_writebuf(curs->level[i].prev_buf_p, 0);
+ libxfs_buf_dirty(curs->level[i].prev_buf_p, 0);
+ libxfs_buf_relse(curs->level[i].prev_buf_p);
}
- libxfs_writebuf(curs->level[i].buf_p, 0);
+ libxfs_buf_dirty(curs->level[i].buf_p, 0);
+ libxfs_buf_relse(curs->level[i].buf_p);
}
}
@@ -681,7 +683,8 @@ prop_freespace_cursor(xfs_mount_t *mp, xfs_agnumber_t agno,
#endif
if (lptr->prev_agbno != NULLAGBLOCK) {
ASSERT(lptr->prev_buf_p != NULL);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_agbno = lptr->agbno;;
lptr->prev_buf_p = lptr->buf_p;
@@ -870,7 +873,8 @@ build_freespace_tree(xfs_mount_t *mp, xfs_agnumber_t agno,
lptr->prev_agbno);
#endif
ASSERT(lptr->prev_agbno != NULLAGBLOCK);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_buf_p = lptr->buf_p;
lptr->prev_agbno = lptr->agbno;
@@ -1046,7 +1050,8 @@ prop_ino_cursor(xfs_mount_t *mp, xfs_agnumber_t agno, bt_status_t *btree_curs,
#endif
if (lptr->prev_agbno != NULLAGBLOCK) {
ASSERT(lptr->prev_buf_p != NULL);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_agbno = lptr->agbno;;
lptr->prev_buf_p = lptr->buf_p;
@@ -1137,7 +1142,8 @@ build_agi(xfs_mount_t *mp, xfs_agnumber_t agno, bt_status_t *btree_curs,
agi->agi_free_level = cpu_to_be32(finobt_curs->num_levels);
}
- libxfs_writebuf(agi_buf, 0);
+ libxfs_buf_dirty(agi_buf, 0);
+ libxfs_buf_relse(agi_buf);
}
/*
@@ -1299,7 +1305,8 @@ build_ino_tree(xfs_mount_t *mp, xfs_agnumber_t agno,
lptr->prev_agbno);
#endif
ASSERT(lptr->prev_agbno != NULLAGBLOCK);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_buf_p = lptr->buf_p;
lptr->prev_agbno = lptr->agbno;
@@ -1451,7 +1458,8 @@ prop_rmap_cursor(
#endif
if (lptr->prev_agbno != NULLAGBLOCK) {
ASSERT(lptr->prev_buf_p != NULL);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_agbno = lptr->agbno;
lptr->prev_buf_p = lptr->buf_p;
@@ -1661,7 +1669,8 @@ _("Insufficient memory to construct reverse-map cursor."));
lptr->prev_agbno);
#endif
ASSERT(lptr->prev_agbno != NULLAGBLOCK);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_buf_p = lptr->buf_p;
lptr->prev_agbno = lptr->agbno;
@@ -1801,7 +1810,8 @@ prop_refc_cursor(
#endif
if (lptr->prev_agbno != NULLAGBLOCK) {
ASSERT(lptr->prev_buf_p != NULL);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_agbno = lptr->agbno;
lptr->prev_buf_p = lptr->buf_p;
@@ -1954,7 +1964,8 @@ _("Insufficient memory to construct refcount cursor."));
lptr->prev_agbno);
#endif
ASSERT(lptr->prev_agbno != NULLAGBLOCK);
- libxfs_writebuf(lptr->prev_buf_p, 0);
+ libxfs_buf_dirty(lptr->prev_buf_p, 0);
+ libxfs_buf_relse(lptr->prev_buf_p);
}
lptr->prev_buf_p = lptr->buf_p;
lptr->prev_agbno = lptr->agbno;
@@ -2142,7 +2153,8 @@ _("Insufficient memory saving lost blocks.\n"));
agf->agf_flcount = 0;
}
- libxfs_writebuf(agfl_buf, 0);
+ libxfs_buf_dirty(agfl_buf, 0);
+ libxfs_buf_relse(agfl_buf);
ext_ptr = findbiggest_bcnt_extent(agno);
agf->agf_longest = cpu_to_be32((ext_ptr != NULL) ?
@@ -2155,7 +2167,8 @@ _("Insufficient memory saving lost blocks.\n"));
ASSERT(be32_to_cpu(agf->agf_refcount_root) !=
be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNTi]));
- libxfs_writebuf(agf_buf, 0);
+ libxfs_buf_dirty(agf_buf, 0);
+ libxfs_buf_relse(agf_buf);
/*
* now fix up the free list appropriately
@@ -2189,7 +2202,8 @@ sync_sb(xfs_mount_t *mp)
update_sb_version(mp);
libxfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
- libxfs_writebuf(bp, 0);
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
}
/*
diff --git a/repair/rmap.c b/repair/rmap.c
index bc53e6c0..c8c851a6 100644
--- a/repair/rmap.c
+++ b/repair/rmap.c
@@ -1231,7 +1231,8 @@ _("setting reflink flag on inode %"PRIu64"\n"),
else
dino->di_flags2 &= cpu_to_be64(~XFS_DIFLAG2_REFLINK);
libxfs_dinode_calc_crc(mp, dino);
- libxfs_writebuf(buf, 0);
+ libxfs_buf_dirty(buf, 0);
+ libxfs_buf_relse(buf);
return 0;
}
diff --git a/repair/scan.c b/repair/scan.c
index f4e4fef5..8b91b27e 100644
--- a/repair/scan.c
+++ b/repair/scan.c
@@ -151,8 +151,10 @@ scan_lbtree(
ASSERT(dirty == 0 || (dirty && !no_modify));
- if ((dirty || badcrc) && !no_modify)
- libxfs_writebuf(bp, 0);
+ if ((dirty || badcrc) && !no_modify) {
+ libxfs_buf_dirty(bp, 0);
+ libxfs_buf_relse(bp);
+ }
else
libxfs_buf_relse(bp);
@@ -2429,13 +2431,17 @@ scan_ag(
sb_dirty += (sbbuf->b_error == -EFSBADCRC);
}
- if (agi_dirty && !no_modify)
- libxfs_writebuf(agibuf, 0);
+ if (agi_dirty && !no_modify) {
+ libxfs_buf_dirty(agibuf, 0);
+ libxfs_buf_relse(agibuf);
+ }
else
libxfs_buf_relse(agibuf);
- if (agf_dirty && !no_modify)
- libxfs_writebuf(agfbuf, 0);
+ if (agf_dirty && !no_modify) {
+ libxfs_buf_dirty(agfbuf, 0);
+ libxfs_buf_relse(agfbuf);
+ }
else
libxfs_buf_relse(agfbuf);
@@ -2443,7 +2449,8 @@ scan_ag(
if (agno == 0)
memcpy(&mp->m_sb, sb, sizeof(xfs_sb_t));
libxfs_sb_to_disk(XFS_BUF_TO_SBP(sbbuf), sb);
- libxfs_writebuf(sbbuf, 0);
+ libxfs_buf_dirty(sbbuf, 0);
+ libxfs_buf_relse(sbbuf);
} else
libxfs_buf_relse(sbbuf);
free(sb);
diff --git a/repair/xfs_repair.c b/repair/xfs_repair.c
index 8642c5cd..2189c9fd 100644
--- a/repair/xfs_repair.c
+++ b/repair/xfs_repair.c
@@ -1126,7 +1126,8 @@ _("Note - stripe unit (%d) and width (%d) were copied from a backup superblock.\
be32_to_cpu(dsb->sb_unit), be32_to_cpu(dsb->sb_width));
}
- libxfs_writebuf(sbp, 0);
+ libxfs_buf_dirty(sbp, 0);
+ libxfs_buf_relse(sbp);
/*
* Done. Flush all cached buffers and inodes first to ensure all
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 14/18] libxfs: remove libxfs_writebuf_int
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (12 preceding siblings ...)
2020-02-20 1:43 ` [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion Darrick J. Wong
@ 2020-02-20 1:44 ` Darrick J. Wong
2020-02-21 14:53 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr Darrick J. Wong
` (3 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:44 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
This function is the same as libxfs_buf_dirty so use that instead.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 1 -
libxfs/rdwr.c | 13 -------------
libxfs/trans.c | 2 +-
3 files changed, 1 insertion(+), 15 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index ad8acc1e..1f6e6d97 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -219,7 +219,6 @@ extern int libxfs_bcache_overflowed(void);
extern xfs_buf_t *libxfs_getbufr(struct xfs_buftarg *, xfs_daddr_t, int);
extern void libxfs_putbufr(xfs_buf_t *);
-extern int libxfs_writebuf_int(xfs_buf_t *, int);
int libxfs_bwrite(struct xfs_buf *);
extern int libxfs_readbufr(struct xfs_buftarg *, xfs_daddr_t, xfs_buf_t *, int, int);
extern int libxfs_readbufr_map(struct xfs_buftarg *, struct xfs_buf *, int);
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index af363bef..9302a698 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -1225,19 +1225,6 @@ libxfs_bwrite(
return bp->b_error;
}
-int
-libxfs_writebuf_int(xfs_buf_t *bp, int flags)
-{
- /*
- * Clear any error hanging over from reading the buffer. This prevents
- * subsequent reads after this write from seeing stale errors.
- */
- bp->b_error = 0;
- bp->b_flags &= ~LIBXFS_B_STALE;
- bp->b_flags |= (LIBXFS_B_DIRTY | flags);
- return 0;
-}
-
/*
* Mark a buffer dirty. The dirty data will be written out when the cache
* is flushed (or at release time if the buffer is uncached).
diff --git a/libxfs/trans.c b/libxfs/trans.c
index 4c208422..e23ae598 100644
--- a/libxfs/trans.c
+++ b/libxfs/trans.c
@@ -863,7 +863,7 @@ buf_item_done(
hold = (bip->bli_flags & XFS_BLI_HOLD);
if (bip->bli_flags & XFS_BLI_DIRTY)
- libxfs_writebuf_int(bp, 0);
+ libxfs_buf_dirty(bp, 0);
bip->bli_flags &= ~XFS_BLI_HOLD;
xfs_buf_item_put(bip);
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (13 preceding siblings ...)
2020-02-20 1:44 ` [PATCH 14/18] libxfs: remove libxfs_writebuf_int Darrick J. Wong
@ 2020-02-20 1:44 ` Darrick J. Wong
2020-02-21 14:53 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 16/18] libxfs: hide libxfs_getbuf_flags Darrick J. Wong
` (2 subsequent siblings)
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:44 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Hide these two functions because they're not used outside of rdwr.c.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 3 ---
libxfs/rdwr.c | 8 ++++++--
2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 1f6e6d97..7d96c2a3 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -216,9 +216,6 @@ extern void libxfs_bcache_flush(void);
extern int libxfs_bcache_overflowed(void);
/* Buffer (Raw) Interfaces */
-extern xfs_buf_t *libxfs_getbufr(struct xfs_buftarg *, xfs_daddr_t, int);
-extern void libxfs_putbufr(xfs_buf_t *);
-
int libxfs_bwrite(struct xfs_buf *);
extern int libxfs_readbufr(struct xfs_buftarg *, xfs_daddr_t, xfs_buf_t *, int, int);
extern int libxfs_readbufr_map(struct xfs_buftarg *, struct xfs_buf *, int);
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 9302a698..bab70dfd 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -52,6 +52,10 @@
* propagation of stale errors into future buffer operations.
*/
+static struct xfs_buf *libxfs_getbufr(struct xfs_buftarg *btp,
+ xfs_daddr_t daddr, int len);
+static void libxfs_putbufr(struct xfs_buf *bp);
+
#define BDSTRAT_SIZE (256 * 1024)
#define IO_BCOMPARE_CHECK
@@ -666,7 +670,7 @@ __libxfs_getbufr(int blen)
return bp;
}
-xfs_buf_t *
+static xfs_buf_t *
libxfs_getbufr(struct xfs_buftarg *btp, xfs_daddr_t blkno, int bblen)
{
xfs_buf_t *bp;
@@ -1360,7 +1364,7 @@ libxfs_bflush(
return bp->b_error;
}
-void
+static void
libxfs_putbufr(xfs_buf_t *bp)
{
if (bp->b_flags & LIBXFS_B_DIRTY)
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 16/18] libxfs: hide libxfs_getbuf_flags
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (14 preceding siblings ...)
2020-02-20 1:44 ` [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr Darrick J. Wong
@ 2020-02-20 1:44 ` Darrick J. Wong
2020-02-21 14:54 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 17/18] libxfs: rename libxfs_readbuf_map to libxfs_buf_read_map Darrick J. Wong
2020-02-20 1:44 ` [PATCH 18/18] libxfs: rename libxfs_getbuf_map to libxfs_buf_get_map Darrick J. Wong
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:44 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Hide this function since it's internal to rdwr.c.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 7 -------
libxfs/rdwr.c | 21 ++++++---------------
2 files changed, 6 insertions(+), 22 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 7d96c2a3..32f8fde7 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -147,9 +147,6 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_getbuf_map(dev, map, nmaps, flags) \
libxfs_trace_getbuf_map(__FUNCTION__, __FILE__, __LINE__, \
(dev), (map), (nmaps), (flags))
-#define libxfs_getbuf_flags(dev, daddr, len, flags) \
- libxfs_trace_getbuf_flags(__FUNCTION__, __FILE__, __LINE__, \
- (dev), (daddr), (len), (flags))
#define libxfs_buf_relse(buf) \
libxfs_trace_putbuf(__FUNCTION__, __FILE__, __LINE__, (buf))
@@ -166,8 +163,6 @@ struct xfs_buf *libxfs_trace_getbuf(const char *func, const char *file,
size_t len);
extern xfs_buf_t *libxfs_trace_getbuf_map(const char *, const char *, int,
struct xfs_buftarg *, struct xfs_buf_map *, int, int);
-extern xfs_buf_t *libxfs_trace_getbuf_flags(const char *, const char *, int,
- struct xfs_buftarg *, xfs_daddr_t, int, unsigned int);
extern void libxfs_trace_putbuf (const char *, const char *, int,
xfs_buf_t *);
@@ -178,8 +173,6 @@ extern xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
void libxfs_buf_dirty(struct xfs_buf *bp, int flags);
extern xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *,
struct xfs_buf_map *, int, int);
-extern xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t,
- int, unsigned int);
void libxfs_buf_relse(struct xfs_buf *);
static inline struct xfs_buf*
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index bab70dfd..f46787a6 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -388,7 +388,6 @@ libxfs_log_header(
#undef libxfs_readbuf_map
#undef libxfs_writebuf
#undef libxfs_getbuf_map
-#undef libxfs_getbuf_flags
xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
@@ -397,8 +396,6 @@ struct xfs_buf *libxfs_buf_get(struct xfs_buftarg *btp, xfs_daddr_t daddr,
size_t len);
xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int);
-xfs_buf_t *libxfs_getbuf_flags(struct xfs_buftarg *, xfs_daddr_t, int,
- unsigned int);
void libxfs_buf_relse(struct xfs_buf *);
#define __add_trace(bp, func, file, line) \
@@ -468,15 +465,6 @@ libxfs_trace_getbuf_map(const char *func, const char *file, int line,
return bp;
}
-xfs_buf_t *
-libxfs_trace_getbuf_flags(const char *func, const char *file, int line,
- struct xfs_buftarg *btp, xfs_daddr_t blkno, int len, unsigned int flags)
-{
- xfs_buf_t *bp = libxfs_getbuf_flags(btp, blkno, len, flags);
- __add_trace(bp, func, file, line);
- return bp;
-}
-
void
libxfs_trace_putbuf(const char *func, const char *file, int line, xfs_buf_t *bp)
{
@@ -780,9 +768,12 @@ __cache_lookup(struct xfs_bufkey *key, unsigned int flags)
return NULL;
}
-struct xfs_buf *
-libxfs_getbuf_flags(struct xfs_buftarg *btp, xfs_daddr_t blkno, int len,
- unsigned int flags)
+static struct xfs_buf *
+libxfs_getbuf_flags(
+ struct xfs_buftarg *btp,
+ xfs_daddr_t blkno,
+ int len,
+ unsigned int flags)
{
struct xfs_bufkey key = {NULL};
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 17/18] libxfs: rename libxfs_readbuf_map to libxfs_buf_read_map
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (15 preceding siblings ...)
2020-02-20 1:44 ` [PATCH 16/18] libxfs: hide libxfs_getbuf_flags Darrick J. Wong
@ 2020-02-20 1:44 ` Darrick J. Wong
2020-02-21 14:54 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 18/18] libxfs: rename libxfs_getbuf_map to libxfs_buf_get_map Darrick J. Wong
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:44 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Rename this function to match the kernel function.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
db/io.c | 2 +-
libxfs/libxfs_io.h | 6 +++---
libxfs/rdwr.c | 8 ++++----
libxfs/trans.c | 4 ++--
repair/da_util.c | 2 +-
5 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/db/io.c b/db/io.c
index 7c7a4624..b81e9969 100644
--- a/db/io.c
+++ b/db/io.c
@@ -542,7 +542,7 @@ set_cur(
if (!iocur_top->bbmap)
return;
memcpy(iocur_top->bbmap, bbmap, sizeof(struct bbmap));
- bp = libxfs_readbuf_map(mp->m_ddev_targp, bbmap->b,
+ bp = libxfs_buf_read_map(mp->m_ddev_targp, bbmap->b,
bbmap->nmaps, 0, ops);
} else {
bp = libxfs_buf_read(mp->m_ddev_targp, blknum, len, 0, ops);
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 32f8fde7..8e9af208 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -135,7 +135,7 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_buf_read(dev, daddr, len, flags, ops) \
libxfs_trace_readbuf(__FUNCTION__, __FILE__, __LINE__, \
(dev), (daddr), (len), (flags), (ops))
-#define libxfs_readbuf_map(dev, map, nmaps, flags, ops) \
+#define libxfs_buf_read_map(dev, map, nmaps, flags, ops) \
libxfs_trace_readbuf_map(__FUNCTION__, __FILE__, __LINE__, \
(dev), (map), (nmaps), (flags), (ops))
#define libxfs_buf_dirty(buf, flags) \
@@ -168,7 +168,7 @@ extern void libxfs_trace_putbuf (const char *, const char *, int,
#else
-extern xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
+extern xfs_buf_t *libxfs_buf_read_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
void libxfs_buf_dirty(struct xfs_buf *bp, int flags);
extern xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *,
@@ -196,7 +196,7 @@ libxfs_buf_read(
{
DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
- return libxfs_readbuf_map(target, &map, 1, flags, ops);
+ return libxfs_buf_read_map(target, &map, 1, flags, ops);
}
#endif /* XFS_BUF_TRACING */
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index f46787a6..531f24e3 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -385,11 +385,11 @@ libxfs_log_header(
#ifdef XFS_BUF_TRACING
-#undef libxfs_readbuf_map
+#undef libxfs_buf_read_map
#undef libxfs_writebuf
#undef libxfs_getbuf_map
-xfs_buf_t *libxfs_readbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
+xfs_buf_t *libxfs_buf_read_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
int libxfs_writebuf(xfs_buf_t *, int);
struct xfs_buf *libxfs_buf_get(struct xfs_buftarg *btp, xfs_daddr_t daddr,
@@ -421,7 +421,7 @@ libxfs_trace_readbuf(
struct xfs_buf *bp;
DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
- bp = libxfs_readbuf_map(btp, &map, 1, flags, ops);
+ bp = libxfs_buf_read_map(btp, &map, 1, flags, ops);
__add_trace(bp, func, file, line);
return bp;
}
@@ -1026,7 +1026,7 @@ libxfs_readbufr_map(struct xfs_buftarg *btp, struct xfs_buf *bp, int flags)
}
struct xfs_buf *
-libxfs_readbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map, int nmaps,
+libxfs_buf_read_map(struct xfs_buftarg *btp, struct xfs_buf_map *map, int nmaps,
int flags, const struct xfs_buf_ops *ops)
{
struct xfs_buf *bp;
diff --git a/libxfs/trans.c b/libxfs/trans.c
index e23ae598..f532e3d6 100644
--- a/libxfs/trans.c
+++ b/libxfs/trans.c
@@ -507,7 +507,7 @@ libxfs_trans_read_buf_map(
*bpp = NULL;
if (tp == NULL) {
- bp = libxfs_readbuf_map(target, map, nmaps, flags, ops);
+ bp = libxfs_buf_read_map(target, map, nmaps, flags, ops);
if (!bp) {
return (flags & XBF_TRYLOCK) ? -EAGAIN : -ENOMEM;
}
@@ -526,7 +526,7 @@ libxfs_trans_read_buf_map(
goto done;
}
- bp = libxfs_readbuf_map(target, map, nmaps, flags, ops);
+ bp = libxfs_buf_read_map(target, map, nmaps, flags, ops);
if (!bp) {
return (flags & XBF_TRYLOCK) ? -EAGAIN : -ENOMEM;
}
diff --git a/repair/da_util.c b/repair/da_util.c
index d1e17ec3..ed2ec3ba 100644
--- a/repair/da_util.c
+++ b/repair/da_util.c
@@ -64,7 +64,7 @@ da_read_buf(
map[i].bm_bn = XFS_FSB_TO_DADDR(mp, bmp[i].startblock);
map[i].bm_len = XFS_FSB_TO_BB(mp, bmp[i].blockcount);
}
- bp = libxfs_readbuf_map(mp->m_dev, map, nex, 0, ops);
+ bp = libxfs_buf_read_map(mp->m_dev, map, nex, 0, ops);
if (map != map_array)
free(map);
return bp;
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [PATCH 18/18] libxfs: rename libxfs_getbuf_map to libxfs_buf_get_map
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
` (16 preceding siblings ...)
2020-02-20 1:44 ` [PATCH 17/18] libxfs: rename libxfs_readbuf_map to libxfs_buf_read_map Darrick J. Wong
@ 2020-02-20 1:44 ` Darrick J. Wong
2020-02-21 14:55 ` Christoph Hellwig
17 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-20 1:44 UTC (permalink / raw)
To: sandeen, darrick.wong; +Cc: linux-xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Rename this function to match the kernel function.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/libxfs_io.h | 6 +++---
libxfs/rdwr.c | 16 ++++++++--------
libxfs/trans.c | 4 ++--
repair/prefetch.c | 2 +-
4 files changed, 14 insertions(+), 14 deletions(-)
diff --git a/libxfs/libxfs_io.h b/libxfs/libxfs_io.h
index 8e9af208..2a451ab2 100644
--- a/libxfs/libxfs_io.h
+++ b/libxfs/libxfs_io.h
@@ -144,7 +144,7 @@ extern struct cache_operations libxfs_bcache_operations;
#define libxfs_buf_get(dev, daddr, len) \
libxfs_trace_getbuf(__FUNCTION__, __FILE__, __LINE__, \
(dev), (daddr), (len))
-#define libxfs_getbuf_map(dev, map, nmaps, flags) \
+#define libxfs_buf_get_map(dev, map, nmaps, flags) \
libxfs_trace_getbuf_map(__FUNCTION__, __FILE__, __LINE__, \
(dev), (map), (nmaps), (flags))
#define libxfs_buf_relse(buf) \
@@ -171,7 +171,7 @@ extern void libxfs_trace_putbuf (const char *, const char *, int,
extern xfs_buf_t *libxfs_buf_read_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
void libxfs_buf_dirty(struct xfs_buf *bp, int flags);
-extern xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *,
+extern xfs_buf_t *libxfs_buf_get_map(struct xfs_buftarg *,
struct xfs_buf_map *, int, int);
void libxfs_buf_relse(struct xfs_buf *);
@@ -183,7 +183,7 @@ libxfs_buf_get(
{
DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
- return libxfs_getbuf_map(target, &map, 1, 0);
+ return libxfs_buf_get_map(target, &map, 1, 0);
}
static inline struct xfs_buf*
diff --git a/libxfs/rdwr.c b/libxfs/rdwr.c
index 531f24e3..52674559 100644
--- a/libxfs/rdwr.c
+++ b/libxfs/rdwr.c
@@ -387,14 +387,14 @@ libxfs_log_header(
#undef libxfs_buf_read_map
#undef libxfs_writebuf
-#undef libxfs_getbuf_map
+#undef libxfs_buf_get_map
xfs_buf_t *libxfs_buf_read_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int, const struct xfs_buf_ops *);
int libxfs_writebuf(xfs_buf_t *, int);
struct xfs_buf *libxfs_buf_get(struct xfs_buftarg *btp, xfs_daddr_t daddr,
size_t len);
-xfs_buf_t *libxfs_getbuf_map(struct xfs_buftarg *, struct xfs_buf_map *,
+xfs_buf_t *libxfs_buf_get_map(struct xfs_buftarg *, struct xfs_buf_map *,
int, int);
void libxfs_buf_relse(struct xfs_buf *);
@@ -450,7 +450,7 @@ libxfs_trace_getbuf(
struct xfs_buf *bp;
DEFINE_SINGLE_BUF_MAP(map, blkno, numblks);
- bp = libxfs_getbuf_map(target, &map, 1, 0);
+ bp = libxfs_buf_get_map(target, &map, 1, 0);
__add_trace(bp, func, file, line);
return bp;
}
@@ -460,7 +460,7 @@ libxfs_trace_getbuf_map(const char *func, const char *file, int line,
struct xfs_buftarg *btp, struct xfs_buf_map *map, int nmaps,
int flags)
{
- xfs_buf_t *bp = libxfs_getbuf_map(btp, map, nmaps, flags);
+ xfs_buf_t *bp = libxfs_buf_get_map(btp, map, nmaps, flags);
__add_trace(bp, func, file, line);
return bp;
}
@@ -805,7 +805,7 @@ reset_buf_state(
}
static struct xfs_buf *
-__libxfs_getbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
+__libxfs_buf_get_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
int nmaps, int flags)
{
struct xfs_bufkey key = {NULL};
@@ -827,12 +827,12 @@ __libxfs_getbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
}
struct xfs_buf *
-libxfs_getbuf_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
+libxfs_buf_get_map(struct xfs_buftarg *btp, struct xfs_buf_map *map,
int nmaps, int flags)
{
struct xfs_buf *bp;
- bp = __libxfs_getbuf_map(btp, map, nmaps, flags);
+ bp = __libxfs_buf_get_map(btp, map, nmaps, flags);
reset_buf_state(bp);
return bp;
}
@@ -1036,7 +1036,7 @@ libxfs_buf_read_map(struct xfs_buftarg *btp, struct xfs_buf_map *map, int nmaps,
return libxfs_readbuf(btp, map[0].bm_bn, map[0].bm_len,
flags, ops);
- bp = __libxfs_getbuf_map(btp, map, nmaps, 0);
+ bp = __libxfs_buf_get_map(btp, map, nmaps, 0);
if (!bp)
return NULL;
diff --git a/libxfs/trans.c b/libxfs/trans.c
index f532e3d6..f4ce23a7 100644
--- a/libxfs/trans.c
+++ b/libxfs/trans.c
@@ -427,7 +427,7 @@ libxfs_trans_get_buf_map(
struct xfs_buf_log_item *bip;
if (!tp)
- return libxfs_getbuf_map(target, map, nmaps, 0);
+ return libxfs_buf_get_map(target, map, nmaps, 0);
/*
* If we find the buffer in the cache with this transaction
@@ -445,7 +445,7 @@ libxfs_trans_get_buf_map(
return bp;
}
- bp = libxfs_getbuf_map(target, map, nmaps, 0);
+ bp = libxfs_buf_get_map(target, map, nmaps, 0);
if (bp == NULL) {
return NULL;
}
diff --git a/repair/prefetch.c b/repair/prefetch.c
index a3858f9a..7f705cc0 100644
--- a/repair/prefetch.c
+++ b/repair/prefetch.c
@@ -120,7 +120,7 @@ pf_queue_io(
* the lock holder is either reading it from disk himself or
* completely overwriting it this behaviour is perfectly fine.
*/
- bp = libxfs_getbuf_map(mp->m_dev, map, nmaps, LIBXFS_GETBUF_TRYLOCK);
+ bp = libxfs_buf_get_map(mp->m_dev, map, nmaps, LIBXFS_GETBUF_TRYLOCK);
if (!bp)
return;
^ permalink raw reply related [flat|nested] 44+ messages in thread
* Re: [PATCH 01/18] libxfs: clean up readbuf flags
2020-02-20 1:42 ` [PATCH 01/18] libxfs: clean up readbuf flags Darrick J. Wong
@ 2020-02-21 14:42 ` Christoph Hellwig
2020-02-21 15:55 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:42 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:42:40PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Create a separate namespace for libxfs_readbuf() flags so that it's a
> little more obvious when we're trying to use the "read or die" logic.
Can we just kill this damn flag instead? Life would be much simpler
if the exit simply moved to the caller. It also kills the exit call
in a library anti-pattern (although of course due to being conditional
it isn't as bad as the real antipattern from the X11 libraries..)
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse
2020-02-20 1:43 ` [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse Darrick J. Wong
@ 2020-02-21 14:43 ` Christoph Hellwig
2020-02-21 15:01 ` Eric Sandeen
1 sibling, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:43 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:00PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Change all the libxfs_putbuf calls to libxfs_buf_relse to match the
> kernel interface, since one is a #define of the other.
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 05/18] libxfs: replace libxfs_getbuf with libxfs_buf_get
2020-02-20 1:43 ` [PATCH 05/18] libxfs: replace libxfs_getbuf with libxfs_buf_get Darrick J. Wong
@ 2020-02-21 14:43 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:43 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:06PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Change all the libxfs_getbuf calls to libxfs_buf_get to match the
> kernel interface, since one is a #define of the other.
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 06/18] libxfs: replace libxfs_readbuf with libxfs_buf_read
2020-02-20 1:43 ` [PATCH 06/18] libxfs: replace libxfs_readbuf with libxfs_buf_read Darrick J. Wong
@ 2020-02-21 14:44 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:44 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:13PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Change all the libxfs_readbuf calls to libxfs_buf_read to match the
> kernel interface. This enables us to hide libxfs_readbuf and simplify
> the userspace buffer interface further.
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does
2020-02-20 1:43 ` [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does Darrick J. Wong
@ 2020-02-21 14:45 ` Christoph Hellwig
2020-02-21 20:43 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:45 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:19PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Make libxfs_bwrite equivalent to libxfs_writebufr. This makes it so
> that libxfs bwrite calls write the buffer to disk immediately and
> without double-freeing the buffer. However, we choose to consolidate
> around the libxfs_bwrite name to match the kernel.
The commit message looks a little odd. Also why is the subject a
different style than the previous messages?
Also if there was a double free shouldn't we fix that in a small
backportable patch first (not that I see such a fix anywhere)?
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 08/18] libxfs: make libxfs_readbuf_verify return an error code
2020-02-20 1:43 ` [PATCH 08/18] libxfs: make libxfs_readbuf_verify return an error code Darrick J. Wong
@ 2020-02-21 14:46 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:46 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:25PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Return the bp->b_error from libxfs_readbuf_verify instead of making
> callers check it.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 09/18] libxfs: make libxfs_readbufr stash the error value in b_error
2020-02-20 1:43 ` [PATCH 09/18] libxfs: make libxfs_readbufr stash the error value in b_error Darrick J. Wong
@ 2020-02-21 14:46 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:46 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:31PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Make libxfs_readbufr stash the error value in b_error, which will make
> the behavior consistent between regular and multi-mapping buffers.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached
2020-02-20 1:43 ` [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached Darrick J. Wong
@ 2020-02-21 14:48 ` Christoph Hellwig
2020-02-21 20:50 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:48 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:37PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Introduce an uncached read function so that userspace can handle them in
> the same way as the kernel. This also eliminates the need for some of
> the libxfs_purgebuf calls (and two trips into the cache code).
>
> Refactor the get/read uncached buffer functions to hide the details of
> uncached buffer-ism in rdwr.c.
Split this into one patch adding the functionality to libxfs and
one each to convert db and copy over with a good rationale for the
switch in each case?
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them
2020-02-20 1:43 ` [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them Darrick J. Wong
@ 2020-02-21 14:49 ` Christoph Hellwig
2020-02-21 15:51 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:49 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:43PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Use the new uncached buffer functions to manage buffers instead of
> open-coding the logic.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Isn't xfs_log_recover.c supposed to be partially shared with the kernel?
I guess I need to port over my changes to the buffer handling there,
which would mean implementing xfs_rw_bdev using preadv/pwritev.
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes
2020-02-20 1:43 ` [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes Darrick J. Wong
@ 2020-02-21 14:51 ` Christoph Hellwig
2020-02-21 15:52 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:51 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:49PM -0800, Darrick J. Wong wrote:
> +/* Prepare an uncached buffer, ready to write something out. */
> +static inline struct xfs_buf *
> +get_write_buf(
> + struct xfs_buftarg *btp,
> + xfs_daddr_t daddr,
> + int bblen)
> +{
> + struct xfs_buf *bp;
> +
> + bp = libxfs_buf_get_uncached(btp, bblen, 0);
> + bp->b_bn = daddr;
> + bp->b_maps[0].bm_bn = daddr;
> + return bp;
> +}
I'd call this alloc_write_buf.
But the patch itself looks good:
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion
2020-02-20 1:43 ` [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion Darrick J. Wong
@ 2020-02-21 14:53 ` Christoph Hellwig
2020-02-21 15:54 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:53 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:43:55PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> libxfs_writebuf is not a well named function -- it marks the buffer
> dirty and then releases the caller's reference. The actual write comes
> when the cache is flushed, either because someone explicitly told the
> cache to flush or because we started buffer reclaim.
>
> Make the buffer release explicit in the callers and rename the function
> to say what it actually does -- it marks the buffer dirty outside of
> transaction context.
FYI, before we made all buffer writes delwri or sync in the kernel the
equivalent function was called xfs_buf_bawrite. But I'm not sure that
was a better name. But maybe libxfs_buf_mark_dirty is a tad better?
Otherwise looks good.
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 14/18] libxfs: remove libxfs_writebuf_int
2020-02-20 1:44 ` [PATCH 14/18] libxfs: remove libxfs_writebuf_int Darrick J. Wong
@ 2020-02-21 14:53 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:53 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:44:02PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> This function is the same as libxfs_buf_dirty so use that instead.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr
2020-02-20 1:44 ` [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr Darrick J. Wong
@ 2020-02-21 14:53 ` Christoph Hellwig
2020-02-21 16:00 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:53 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:44:09PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Hide these two functions because they're not used outside of rdwr.c.
Can we also pick a better name? That r postfix always seem very strange
to me.
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 16/18] libxfs: hide libxfs_getbuf_flags
2020-02-20 1:44 ` [PATCH 16/18] libxfs: hide libxfs_getbuf_flags Darrick J. Wong
@ 2020-02-21 14:54 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:54 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:44:15PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Hide this function since it's internal to rdwr.c.
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 17/18] libxfs: rename libxfs_readbuf_map to libxfs_buf_read_map
2020-02-20 1:44 ` [PATCH 17/18] libxfs: rename libxfs_readbuf_map to libxfs_buf_read_map Darrick J. Wong
@ 2020-02-21 14:54 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:54 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:44:21PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Rename this function to match the kernel function.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 18/18] libxfs: rename libxfs_getbuf_map to libxfs_buf_get_map
2020-02-20 1:44 ` [PATCH 18/18] libxfs: rename libxfs_getbuf_map to libxfs_buf_get_map Darrick J. Wong
@ 2020-02-21 14:55 ` Christoph Hellwig
0 siblings, 0 replies; 44+ messages in thread
From: Christoph Hellwig @ 2020-02-21 14:55 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: sandeen, linux-xfs
On Wed, Feb 19, 2020 at 05:44:27PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Rename this function to match the kernel function.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Looks good,
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse
2020-02-20 1:43 ` [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse Darrick J. Wong
2020-02-21 14:43 ` Christoph Hellwig
@ 2020-02-21 15:01 ` Eric Sandeen
1 sibling, 0 replies; 44+ messages in thread
From: Eric Sandeen @ 2020-02-21 15:01 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: linux-xfs
On 2/19/20 7:43 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Change all the libxfs_putbuf calls to libxfs_buf_relse to match the
> kernel interface, since one is a #define of the other.
I'll try to properly review, but /thank you/ - I had this patch on my stack
for a while but hadn't ever sent it. This was just dumb noise, good to see
it go.
-Eric
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them
2020-02-21 14:49 ` Christoph Hellwig
@ 2020-02-21 15:51 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 15:51 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:49:59AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:43:43PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Use the new uncached buffer functions to manage buffers instead of
> > open-coding the logic.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
>
> Isn't xfs_log_recover.c supposed to be partially shared with the kernel?
> I guess I need to port over my changes to the buffer handling there,
> which would mean implementing xfs_rw_bdev using preadv/pwritev.
It ought to be, but the file has fallen very far out of sync with the
kernel and I didn't want to launch a log recovery reconstruction
project. This patch is useful to get us to the point of having one
buffer api and using it consistently, but that's about it...
--D
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes
2020-02-21 14:51 ` Christoph Hellwig
@ 2020-02-21 15:52 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 15:52 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:51:01AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:43:49PM -0800, Darrick J. Wong wrote:
> > +/* Prepare an uncached buffer, ready to write something out. */
> > +static inline struct xfs_buf *
> > +get_write_buf(
> > + struct xfs_buftarg *btp,
> > + xfs_daddr_t daddr,
> > + int bblen)
> > +{
> > + struct xfs_buf *bp;
> > +
> > + bp = libxfs_buf_get_uncached(btp, bblen, 0);
> > + bp->b_bn = daddr;
> > + bp->b_maps[0].bm_bn = daddr;
> > + return bp;
> > +}
>
> I'd call this alloc_write_buf.
Agreed, will fix.
--D
> But the patch itself looks good:
>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion
2020-02-21 14:53 ` Christoph Hellwig
@ 2020-02-21 15:54 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 15:54 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:53:05AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:43:55PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > libxfs_writebuf is not a well named function -- it marks the buffer
> > dirty and then releases the caller's reference. The actual write comes
> > when the cache is flushed, either because someone explicitly told the
> > cache to flush or because we started buffer reclaim.
> >
> > Make the buffer release explicit in the callers and rename the function
> > to say what it actually does -- it marks the buffer dirty outside of
> > transaction context.
>
> FYI, before we made all buffer writes delwri or sync in the kernel the
> equivalent function was called xfs_buf_bawrite.
Heh, yeah... I've been reviewing old Unixen books lately and finally
know where all those weird names came from.
> But I'm not sure that
> was a better name. But maybe libxfs_buf_mark_dirty is a tad better?
Ok.
--D
> Otherwise looks good.
>
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 01/18] libxfs: clean up readbuf flags
2020-02-21 14:42 ` Christoph Hellwig
@ 2020-02-21 15:55 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 15:55 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:42:47AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:42:40PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Create a separate namespace for libxfs_readbuf() flags so that it's a
> > little more obvious when we're trying to use the "read or die" logic.
>
> Can we just kill this damn flag instead? Life would be much simpler
> if the exit simply moved to the caller. It also kills the exit call
> in a library anti-pattern (although of course due to being conditional
> it isn't as bad as the real antipattern from the X11 libraries..)
Heh. It was only now that I realized that there are ~8 callers of the
"fail on ioerror" flag. Yes, let's get rid of them both.
--D
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr
2020-02-21 14:53 ` Christoph Hellwig
@ 2020-02-21 16:00 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 16:00 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:53:54AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:44:09PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Hide these two functions because they're not used outside of rdwr.c.
>
> Can we also pick a better name? That r postfix always seem very strange
> to me.
They're odd, but I also don't really have a good suggestion...
getbufr -> xfs_buf_get_raw?
putbufr -> xfs_buf_commit?
--D
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does
2020-02-21 14:45 ` Christoph Hellwig
@ 2020-02-21 20:43 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 20:43 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:45:47AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:43:19PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Make libxfs_bwrite equivalent to libxfs_writebufr. This makes it so
> > that libxfs bwrite calls write the buffer to disk immediately and
> > without double-freeing the buffer. However, we choose to consolidate
> > around the libxfs_bwrite name to match the kernel.
>
> The commit message looks a little odd. Also why is the subject a
> different style than the previous messages?
Ok, changed to:
"libxfs: rename libxfs_writebufr to libxfs_bwrite
"Rename this function so that we have an API that matches the kernel
without a #define."
> Also if there was a double free shouldn't we fix that in a small
> backportable patch first (not that I see such a fix anywhere)?
Hmmm... I can't find one either. I think that's just an anachronism
from an earlier version of this series.
--D
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached
2020-02-21 14:48 ` Christoph Hellwig
@ 2020-02-21 20:50 ` Darrick J. Wong
2020-02-21 20:59 ` Darrick J. Wong
0 siblings, 1 reply; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 20:50 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 06:48:33AM -0800, Christoph Hellwig wrote:
> On Wed, Feb 19, 2020 at 05:43:37PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Introduce an uncached read function so that userspace can handle them in
> > the same way as the kernel. This also eliminates the need for some of
> > the libxfs_purgebuf calls (and two trips into the cache code).
> >
> > Refactor the get/read uncached buffer functions to hide the details of
> > uncached buffer-ism in rdwr.c.
>
> Split this into one patch adding the functionality to libxfs and
> one each to convert db and copy over with a good rationale for the
> switch in each case?
Both programs use the uncached read for the same purpose, which is to
read the superblock without polluting the buffer cache when we haven't
yet established the filesystem sector size.
The only reason why repair doesn't need patching is that reaches into
the buffer cache internals to call lseek and read by hand. I guess that
should be fixed, separately.
--D
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached
2020-02-21 20:50 ` Darrick J. Wong
@ 2020-02-21 20:59 ` Darrick J. Wong
0 siblings, 0 replies; 44+ messages in thread
From: Darrick J. Wong @ 2020-02-21 20:59 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: sandeen, linux-xfs
On Fri, Feb 21, 2020 at 12:50:28PM -0800, Darrick J. Wong wrote:
> On Fri, Feb 21, 2020 at 06:48:33AM -0800, Christoph Hellwig wrote:
> > On Wed, Feb 19, 2020 at 05:43:37PM -0800, Darrick J. Wong wrote:
> > > From: Darrick J. Wong <darrick.wong@oracle.com>
> > >
> > > Introduce an uncached read function so that userspace can handle them in
> > > the same way as the kernel. This also eliminates the need for some of
> > > the libxfs_purgebuf calls (and two trips into the cache code).
> > >
> > > Refactor the get/read uncached buffer functions to hide the details of
> > > uncached buffer-ism in rdwr.c.
> >
> > Split this into one patch adding the functionality to libxfs and
> > one each to convert db and copy over with a good rationale for the
> > switch in each case?
>
> Both programs use the uncached read for the same purpose, which is to
> read the superblock without polluting the buffer cache when we haven't
> yet established the filesystem sector size.
HAH nope, despite the very similiar code, they do it for different
reasons. Separate patches it is.
--D
> The only reason why repair doesn't need patching is that reaches into
> the buffer cache internals to call lseek and read by hand. I guess that
> should be fixed, separately.
>
> --D
^ permalink raw reply [flat|nested] 44+ messages in thread
end of thread, other threads:[~2020-02-21 20:59 UTC | newest]
Thread overview: 44+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-20 1:42 [PATCH 00/18] xfsprogs: refactor buffer function names Darrick J. Wong
2020-02-20 1:42 ` [PATCH 01/18] libxfs: clean up readbuf flags Darrick J. Wong
2020-02-21 14:42 ` Christoph Hellwig
2020-02-21 15:55 ` Darrick J. Wong
2020-02-20 1:42 ` [PATCH 02/18] libxfs: clean up writebuf flags Darrick J. Wong
2020-02-20 1:42 ` [PATCH 03/18] libxfs: remove LIBXFS_EXIT_ON_FAILURE Darrick J. Wong
2020-02-20 1:43 ` [PATCH 04/18] libxfs: replace libxfs_putbuf with libxfs_buf_relse Darrick J. Wong
2020-02-21 14:43 ` Christoph Hellwig
2020-02-21 15:01 ` Eric Sandeen
2020-02-20 1:43 ` [PATCH 05/18] libxfs: replace libxfs_getbuf with libxfs_buf_get Darrick J. Wong
2020-02-21 14:43 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 06/18] libxfs: replace libxfs_readbuf with libxfs_buf_read Darrick J. Wong
2020-02-21 14:44 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 07/18] libxfs: make libxfs_bwrite do what libxfs_writebufr does Darrick J. Wong
2020-02-21 14:45 ` Christoph Hellwig
2020-02-21 20:43 ` Darrick J. Wong
2020-02-20 1:43 ` [PATCH 08/18] libxfs: make libxfs_readbuf_verify return an error code Darrick J. Wong
2020-02-21 14:46 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 09/18] libxfs: make libxfs_readbufr stash the error value in b_error Darrick J. Wong
2020-02-21 14:46 ` Christoph Hellwig
2020-02-20 1:43 ` [PATCH 10/18] libxfs: introduce libxfs_buf_read_uncached Darrick J. Wong
2020-02-21 14:48 ` Christoph Hellwig
2020-02-21 20:50 ` Darrick J. Wong
2020-02-21 20:59 ` Darrick J. Wong
2020-02-20 1:43 ` [PATCH 11/18] libxlog: use uncached buffers instead of open-coding them Darrick J. Wong
2020-02-21 14:49 ` Christoph Hellwig
2020-02-21 15:51 ` Darrick J. Wong
2020-02-20 1:43 ` [PATCH 12/18] libxfs: use uncached buffers for initial mkfs writes Darrick J. Wong
2020-02-21 14:51 ` Christoph Hellwig
2020-02-21 15:52 ` Darrick J. Wong
2020-02-20 1:43 ` [PATCH 13/18] libxfs: straighten out libxfs_writebuf confusion Darrick J. Wong
2020-02-21 14:53 ` Christoph Hellwig
2020-02-21 15:54 ` Darrick J. Wong
2020-02-20 1:44 ` [PATCH 14/18] libxfs: remove libxfs_writebuf_int Darrick J. Wong
2020-02-21 14:53 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 15/18] libxfs: hide libxfs_{get,put}bufr Darrick J. Wong
2020-02-21 14:53 ` Christoph Hellwig
2020-02-21 16:00 ` Darrick J. Wong
2020-02-20 1:44 ` [PATCH 16/18] libxfs: hide libxfs_getbuf_flags Darrick J. Wong
2020-02-21 14:54 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 17/18] libxfs: rename libxfs_readbuf_map to libxfs_buf_read_map Darrick J. Wong
2020-02-21 14:54 ` Christoph Hellwig
2020-02-20 1:44 ` [PATCH 18/18] libxfs: rename libxfs_getbuf_map to libxfs_buf_get_map Darrick J. Wong
2020-02-21 14:55 ` Christoph Hellwig
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.